ANALYZING ELECTRIC VEHICLE DIFFUSION SCENARIOS FOR

Transkript

ANALYZING ELECTRIC VEHICLE DIFFUSION SCENARIOS FOR ISTANBUL
by
Özlem Turan
B.S., Industrial Engineering, Istanbul Technical University, 2011
Submitted to the Institute for Graduate Studies in
Science and Engineering in partial fulfillment of
the requirements for the degree of
Master of Science
Graduate Program in Industrial Engineering
Boğaziçi University
2014
ii
ANALYZING ELECTRIC VEHICLE DIFFUSION SCENARIOS FOR ISTANBUL
APPROVED BY:
Assist. Prof. Gönenç Yücel
………………
(Thesis Supervisor)
Prof. Gürkan Kumbaroğlu
………………
Prof. Aslı Sencer
………………
DATE OF APPROVAL:
iii
to my family
iv
ACKNOWLEDGEMENTS
Foremost, I would like to thank my thesis supervisor Assoc. Prof. Gönenç Yücel for
his orientation, guidance, patience, and support throughout this study. This study would not
have been possible without his support and his tolerance.
I gratefully thank my thesis committee members Prof. Gürkan Kumbaroğlu and Prof.
Aslı Sencer for taking part in my thesis committee and taking time to examine this thesis.
I offer my deepest gratitude to my family for their love, support, and patience,
throughout my life.
I own special acknowledgments to my roommate Yasemin Kalafatoğlu and to my
neighbor Pelin Ekmen for kindly offering their never-ending help, emotional support, and
long tea hours, to Kadir Yıldız for his never-ending help, his funny stories, our endless
plans for establishing a business, and, to Özge Sürer for her sincere help and inspiring
motivation, and to Gizem Bacaksızlar for being ready to listen, her emotional support, and
her traditional sides. I am so lucky for having such great friendships. Thanks to them, my
time here was much more colorful and encouraging than I could ever think of.
I wish to thank İnci Öykü Yener-Roderburg for her great friendship and for her full
support since high school.
Finally, I would like to point out that it is a great pleasure for me to make my thesis
in Boğaziçi University and to be a part of this family.
v
ABSTRACT
ANALYZING ELECTRIC VEHICLE DIFFUSION
SCENARIOS FOR ISTANBUL
In this study, a dynamic simulation model for electric vehicle (EV) diffusion is
constructed. The objective of this work is to investigate two main questions; what are the
plausible diffusion patterns of electric vehicles for Istanbul under different scenarios
developed considering both local and global socio-economic, governmental, technological
factors and their interaction with each other? Secondly, what is the extent of the diffusion
rate that can be expected in Istanbul after three decades? The model is validated by
standard structure and behavior tests. After, various scenario and policy analysis are
performed. The results show that fleet market share of battery electric vehicle (BEV) and
hybrid electric vehicle (HEV) would likely reach around 19.76% and 20.77% respectively
by 2042 in Istanbul. In addition, CO 2 reduction in the transportation sector would only
reach around 17.32% in 2042. Moreover, both gasoline and electricity cost influence EV
diffusion. However, their impact on EV diffusion is mainly related with a mobility cost gap
between gasoline and electricity. Furthermore, technological improvement would lead
BEV sales to increase. However, if battery technology cannot keep pace with or exceed
CV technology, technological improvements would less likely create a significant raise in
the BEV sales. Contrary to expectations, even if no technological improvements were
progressed, BEVs would still likely succeed to penetrate around 10% of the market with its
current technology within the 30 years. Moreover, a sufficient number of recharging points
may lead to faster diffusion of BEV‟s as well, causing higher fleet market share overall.
Both marketing activities and word of mouth have a remarkable impact on rapid EV
diffusion. Besides, increase in repurchasing rate may cause faster EV penetration.
Subsidies would have a small impact on EV sales. Finally, applying the 3% private
consumption tax (PCT) instead of the 37% PCT for HEV may increase HEV sales but does
not show considerable change on HEV sales.
vi
ÖZET
.
ISTANBUL İÇİN ELEKTRİKLİ ARABA YAYILIM
SENARYOLARININ ANALİZİ
Bu çalışmada, elektrikli arabaların (EA) yayılımı için dinamik bir benzetim modeli
kurulmuştur. Çalışmanın amacı, iki önemli soruya araştırmaktır; yerel ve global sosyoekonomik, hükümetsel, teknolojik faktörler ve onların birbirleriyle etkileşimleri göz
önünde bulundurularak geliştirilen senaryolar altında, elektrikli arabaların makul yayılım
davranışları nelerdir? İkincisi ise Istanbul‟da 30 yıl içinde elektrikli arabaların beklenen
yayılım oranı nedir? Model, standart yapısal ve davranışsal testler ile doğrulanmıştır. Daha
sonra, farklı senaryo ve politika analizleri uygulanmıştır. Sonuçlar, 2042 yılında,
İstanbul‟daki bateri elektrikli araba (BEA) ve hibrit elektrikli araba (HEA) yayılımının
sırasıyla 19.76% ve 20.77% civarında olacağını göstermektedir. Ayrıca 2042 yılında,
ulaşım sektöründe, CO2 miktarında 17.32% civarında azalma olacağı görülmüştür. Bunun
dışında, hem benzin hem de elektrik maliyetinin EA difüzyonunu etkilediği, fakat önemli
etkinin aralarındaki maliyet farkına önemli derecede bağlı olduğu görülmüştür. Teknolojik
gelişmeler BEA satışlarını artırmaktadır fakat bateri teknolojisi, konvansiyonel araçların
seviyesine erişmeden ya da onları geçmeden, bu gelişmeler BAE satışlarında çok büyük bir
değişiklik oluşturmazlar. Ayrıca, hiçbir teknolojik gelişme olmasa bile BEA var olan
teknolojiyle pazarın %10‟una nüfuz edebilir. Bunun dışında, yeterli şarj istasyonu sayısı
yayılımı hızlandırmakta ve artırmaktadır. Hem pazarlama aktivitelerinin hem de insanların
EA hakkında bilgi yaymaya yarayan davranışlarının EA yayılımı üzerinde büyük ve
hızlandırıcı bir etkisi vardır. Araba değişim hızının artması daha hızlı bir EA yayılımına
neden olabilir fakat para yardımı stratejileri elektrikli araba satışlarında çok etkili
olmamaktadır. Son olarak hibrit araçlarda, %37 yerine 3% özel tüketim vergisi
kullanılması, araç satışlarını artırabilir ama çok büyük bir değişikliğe neden olmaz.
vii
TABLE OF CONTENTS
ACKNOWLEDGEMENTS .................................................................................................. iv
ABSTRACT........................................................................................................................... v
ÖZET .................................................................................................................................... vi
LIST OF FIGURES ............................................................................................................... x
LIST OF TABLES ............................................................................................................... xv
LIST OF SYMBOLS .......................................................................................................... xvi
LIST OF ACRONYMS/ABBREVIATIONS ....................................................................xvii
1. INTRODUCTION ............................................................................................................. 1
1.1. History of Electric Vehicles ....................................................................................... 4
1.2. Types of Electric Vehicles ......................................................................................... 5
2. PROBLEM DEFINITION AND RESEARCH OBJECTIVES ......................................... 8
3. LITERATURE REVIEW ................................................................................................ 12
4. METHODOLOGY........................................................................................................... 16
5. DESCRIPTION OF THE MODEL ................................................................................. 19
5.1. Vehicles Fleet Sector................................................................................................ 23
5.1.1. Background Information ................................................................................ 23
5.1.2. Description of the Structure ........................................................................... 24
5.1.3. Parameter Estimation and Assumptions ........................................................ 26
5.2. Vehicle Market Sector.............................................................................................. 26
5.3. Customer Perception Sector ..................................................................................... 28
5.3.2.1. Time Utility ……….... ..................................................................... 31
5.3.2.2. Purchase Price Utility.. ..................................................................... 33
5.3.2.3. Operating Cost Utility. ..................................................................... 34
5.3.2.4. Emission Utility… ............................................................................ 35
viii
5.4. Customers Awareness Sector ................................................................................... 38
5.5. Infrastructure Sector ................................................................................................. 43
5.6. Environmental Impact Sector ................................................................................... 46
6. VALIDATION AND ANALYSIS OF THE MODEL .................................................... 49
6.1. Model Validation...................................................................................................... 49
6.1.1. Structural Validity.......................................................................................... 49
6.1.1.1. Direct Structure Tests. ...................................................................... 50
6.1.1.2. Structure Oriented Behavior Tests…................................................ 50
6.1.2. Behavior Validation ....................................................................................... 57
6.2. Analyses of the Base Behavior................................................................................. 58
7. SCENARIO AND POLICY ANALYSIS ........................................................................ 64
7.1. Scenario Analysis ..................................................................................................... 64
7.1.1. Electricity and Gasoline Costs Related Scenarios (Scenario 1)..................... 64
7.1.1.1. Constant Electricity and Gasoline Costs (Scenario 1_1).. ................ 65
7.1.1.2. Low Level of Sensitivity to Electricity Demand (Scenario 1_2)…..67
7.1.1.3. High Level of Sensitivity to Electricity Demand (Scenario 1_3).... . 68
7.1.1.4. High Gasoline Cost vs Normal Electricity Cost (Scenario 1_4).. .... 70
7.1.2. Technological Development Related Scenarios (Scenario 2)........................ 72
7.1.2.1. Moderate Technological Improvement (Scenario 2_1).. .................. 73
7.1.2.2. Optimistic Improvements (Scenario 2_2)…..................................... 75
7.1.2.3. No Improvement (Scenario 2_3).. .................................................... 77
7.1.3. Recharging Infrastructure Based Scenarios (Scenario 3) .............................. 79
7.1.4. Introducing only BEV to the Market (Scenario 4)......................................... 80
7.1.5. Word of Mouth Related Scenarios (WoM) (Scenario 5) ............................... 82
ix
7.1.5.1. Intensive non-EV Drivers‟ Word of Mouth (Scenario 5_1)... .......... 82
7.1.5.2. Intensive EV drivers‟ Word of Mouth (Scenario 5_2).. ................... 84
7.1.6. Repurchasing Rate (Scenario 6)..................................................................... 85
7.2. Policy Analysis......................................................................................................... 87
7.2.1. Subsidy Based Policies (Policy 1) ................................................................. 87
7.2.1.1. 5000 TL Subsidy for BEV (Policy 1_1_1). ...................................... 88
7.2.1.2. 10000TL Subsidy for the First 10 Years for BEV (Policy 1_2_1)... 88
7.2.1.3. 5000 TL Subsidy for HEV (Policy 1_1_2). ...................................... 89
7.2.1.4. 10000TL Subsidy for 10 Years for HEV (Policy 1_2_2). ................ 90
7.2.1.5. 5000 TL Subsidy for Both BEV and HEV (Policy 1_3). ................. 91
7.2.1.6. 10000TL Subsidy for 10 Years for Both EVs (Policy 1_4). ............ 92
7.2.2. Tax Based Policy (Policy 2) .......................................................................... 94
7.2.3. Marketing Based Policies (Policy 3).............................................................. 96
7.2.3.1. No Marketing Activities (Policy 3_1).. ............................................ 97
7.2.3.2. Less Marketing Activities (Policy 3_2) .. ......................................... 98
7.2.3.3. Marketing Activities for Limited Duration (Policy 3_3): ............... 99
7.3. Combination of Scenario and Policies ................................................................... 102
7.3.1. High Electricity Price and Over Marketing Activities (Combination 1) ..... 102
7.3.2. High Gasoline Cost and Bad Recharging Infrastructure (Combination 2) .. 103
7.3.3. Advanced Improvement and No Marketing (Combination 3) ..................... 104
7.3.4. Tax Regulation for HEV and Optimal Progress for BEV (Combination 4) 105
8. CONCLUSION .............................................................................................................. 113
APPENDIX A: MODEL EQUATIONS............................................................................ 117
APPENDIX B: PARAMETER VALUES ......................................................................... 126
APPENDIX C: SENSITIVITY RESULTS ....................................................................... 129
REFERENCES .................................................................................................................. 132
x
LIST OF FIGURES
Figure 4.1.
Causal loop diagram of population model. ............................................... 17
Figure 4.2.
Stock- flow diagram of population model. ................................................ 18
Figure 5.1.
Relations between sectors. ........................................................................ 20
Figure 5.2.
Simplified causal loop diagram. ............................................................... 21
Figure 5.3.
Simplified stock- flow diagram of vehicle fleet sector. ............................ 24
Figure 5.4.
Population projection of Istanbul. ............................................................ 26
Figure 5.5.
Simplified diagram of perceived utility sector. ........................................ 30
Figure 5.6.
Effect function of infrastructure on BEV time utility. ............................. 37
Figure 5.7.
Simplified stock- flow diagram of customer awareness sector. ................ 39
Figure 5.8.
Simplified stock- flow diagram of infrastructure sector. ........................... 44
Figure 6.1.
BEV sales market share under the 1st ect. ................................................ 51
Figure 6.2.
Percentage of potential EV customers under 2 th ect. ............................... 52
Figure 6.3.
BEV share among potential EV customers under 3 rd ect. ........................ 53
Figure 6.4.
HEV share among potential EV customers under 3 rd ect. .......................... 53
Figure 6.5.
CV share among potential EV customers under 3 rd ect. ........................... 54
xi
Figure 6.6.
Sensitivity result for effectiveness of WoM of EV-drivers. ..................... 55
Figure 6.7.
Sensitivity result for motorization rate. .................................................... 56
Figure 6.8.
Sensitivity result for weight of emission utility. ...................................... 57
Figure 6.9.
Fleet share patterns of this study. ............................................................. 58
Figure 6.10.
Fleet share pattern of the work of Wansart and Schnieder. ...................... 58
Figure 6.11.
Sales market share of vehicles under the base run. .................................. 59
Figure 6.12.
Total number of each vehicle under the base run. .................................... 60
Figure 6.13.
Fleet market share of vehicles under the base run. ................................... 61
Figure 6.14.
Percentage of potential EV customers under the base run. ...................... 62
Figure 6.15.
Reduction of CO 2 under the base run. ...................................................... 63
Figure 7.1.
BEV fleet market share under the Scenario 1_1. ...................................... 66
Figure 7.2.
HEV fleet market share under the Scenario 1_1. ..................................... 66
Figure 7.3.
Electricity and gasoline costs under the Scenario 1_2. ............................ 67
Figure 7.4.
BEV fleet market share under the Scenario 1_2. ..................................... 68
Figure 7.5.
Electricity vs gasoline prices under the Scenario 1_3. ............................. 69
Figure 7.6.
Sales market share of BEV and HEV under the Scenario 1_3. ................ 69
Figure 7.7.
xii
Figure 7.8.
Gasoline vs electricity cost under the Scenario 1_4. ................................ 70
Figure 7.9.
Sales market shares of BEV and HEV under the Scenario 1_4. .............. 71
Figure 7.10.
Fleet market shares of BEV and HEV under the Scenario 1_4. ............... 71
Figure 7.11.
BEV driving range under the Scenario 2_1. ............................................. 73
Figure 7.12.
BEV refueling time under the Scenario 2_1. ............................................ 74
Figure 7.13.
BEV maintenance cost under the Scenario 2_1.
Figure 7.14.
Fleet market share of BEV and HEV under the Scenario 2_1 . ................ 75
Figure 7.15.
Technological and infrastructural improvement under the Scenario 2_2. . 76
Figure 7.16.
BEV fleet market share under the the Scenario 2_2. ................................ 76
Figure 7.17.
BEV fleet market share under the Scenario 2_3. ..................................... 77
Figure 7.18.
HEV fleet market share under the Scenario 2_3. ..................................... 78
Figure 7.19.
BEV market share under different infrastructure conditions. .................. 79
Figure 7.20.
Fleet market share under Scenario 4. ....................................................... 81
Figure 7.21.
Reductions of CO 2 under the base run vs Scenario 4. .............................. 81
Figure 7.22.
Figure 7.23.
Fleet market share of BEV and HEV under the Scenario 5_1. ................ 83
Figure 7.24.
..................................... 74
xiii
Figure 7.25.
Fleet market share of BEV and HEV under the Scenario 5_2. ................ 84
Figure 7.26.
BEV fleet market share under the re-purchasing scenario. ...................... 86
Figure 7.27.
HEV fleet market share under the re-purchasing scenario. ...................... 86
Figure 7.28.
Fleet market share of BEV under the Policy 1_1_1. ................................ 88
Figure 7.29.
Fleet market share of BEV under the Policy 1_2_1. ................................ 89
Figure 7.30.
Fleet share of HEV under the Policy 1_1_2. ............................................ 90
Figure 7.31.
Fleet market share of HEV under the Policy 1_2_2. ................................ 91
Figure 7.32.
Fleet market share of BEV and HEV under the Policy 1_3. .................... 92
Figure 7.33.
Figure 7.34.
Sales market share of HEV under the Policy 2. ....................................... 95
Figure 7.35.
Fleet market share of HEV under the Policy 2. ........................................ 95
Figure 7.36.
Sales market share of BEV and HEV under the Policy 3_1. .................... 97
Figure 7.37.
Figure 7.38.
Sales market share of BEV and HEV under the Policy 3_2. .................... 98
Figure 7.39.
Figure 7.40.
Sales market share of BEV under different marketing strategies. .......... 100
Figure 7.41.
Fleet market share of BEV under different marketing strategies. .......... 100
xiv
Figure 7.42.
Sales market share of BEV and HEV under the Combination 1. ........... 103
Figure 7.43.
Figure 7.44.
Figure 7.45.
Sales market share of BEV under the Combination 4. ........................... 106
Figure 7.46.
Sales market share of HEV under the Combination 4. ............................ 106
xv
LIST OF TABLES
Table 1.1.
Comparative properties of vehicle types. .................................................... 7
Table 7.1.
The results of electricity and gasoline costs related scenarios. ................ 72
Table 7.2.
The results of technological development related scenarios. ................... 78
Table 7.3.
The results of recharging infrastructure related scenarios. ....................... 80
Table 7.4.
The results of introducing only BEV to the market. ................................. 82
Table 7.5.
The results of WoM scenarios. ................................................................. 85
Table 7.6.
The results of repurchasing scenarios. ..................................................... 87
Table 7.7.
The results of subsidy based policies......................................................... 93
Table 7.8.
The results of private consumption tax based policy. .............................. 96
Table 7.9.
The results of different marketing policies. ........................................... 101
xvi
LIST OF SYMBOLS
δ,j
Sales market share of i- type vehicle in all customers
δi,j
Sales market share of i- type vehicle which belongs to group j
u i,j
Total perceived utility of vehicle type i which belongs to group
wikj
Weight of utility k for i - type of vehicle which belongs to group j
tui,j
Time utility of i type of vehicle which belongs to group j
α(t)
Value of the attribute at time t
E(t)
Denotes value of cumulative experience at time t
α
Denotes learning factor
nf
Denotes normalization factor
pui,j
Purchase price utility of i type of vehicle which belongs to group j
oui,j
Operating cost utility of i type of vehicle which belongs to group j
rui,j
Refueling cost utility of i type of vehicle which belongs to group j
mui,j
Maintenance cost utility of i type of vehicle which belongs to group j
eui,j
Emission utility of i type of vehicle which belongs to group j
VEV
Total number of EV (sum of BEV and HEV) in Istanbul.
Vt
Total number of vehicle in Istanbul.
xvii
LIST OF ACRONYMS/ABBREVIATIONS
AFV
Alternative Fuel Vehicle
B
Balancing
BEV
Battery electric vehicle
CO 2
Carbon dioxide
CV
Conventional vehicle
Dmnl
dimensionless
ect
extreme condition test
EV
Electric vehicle
HEV
Hybrid electric vehicle
ICE
Internal combustion engine
ICEV
Internal combustion engine vehicle
OECD
Organization for Economic Co-operation and Development
PHEV
Plug- in hybrid electric vehicle
R
Reinforcing
RQ
Research question
SNM
Strategic niche management
sms
Sales market share
TUBITAK
The Scientific and Technological Research Council of Turkey
WoM
Word-of- mouth
1
1. INTRODUCTION
Nowadays vehicles that are powered by internal combustion engines (ICEs), which
transform the chemical energy of fuel to the thermal and mechanical energy, occupy major
role in ground transportation industry all over the world [1]. The transportation industry
produces high amounts of greenhouse gases and pollutant emissions. For this reason,
internal combustion engine vehicles (ICEVs) can be seen as one of the major contributors
to air pollution. ICEVs use petroleum-based fuels that increase the CO 2 level and other
emissions in the air [2]. CO 2 is one of the major greenhouse gases that is emitted to the
atmosphere through burning fossil fuels [3]. Fuglestvedt, Berntsen, Myhre, Rypdal, and
Skeie indicate that 20-25% of the global CO2 emission stems from the transportation sector
which is potential cause of global warming [2]. From 1990 to 2001, different sectors in
European Union reduced their global greenhouse gas emission rates. However, emission
originating from transportation, particularly road transportation, increased about 21%
during the same period [4].
40% of the global energy demand, including almost all of the fuel consumption of
transportation system is supplied by conventional oil [5]. Heated debates started with the
modern oil era in the mid-1800s in relation to the possibility of reaching the peak point of
the global conventional oil 1 . This situation is a potential problem for the urban
transportation system. Since reaching to peak point of oil means facing with a fuel
shortage. In case of reaching to the peak point of oil production, reliance on oil will
generate drastic global challenges in petroleum-based transportation sector such as unmet
demand, high fuel prices, and oil black- market [5,6].
Emergence of environmental and energy related concerns have given birth to
ongoing debates about how world can overcome global warming, air pollution, and limited
oil problem. One of the suggestions is to replace fossil-powered internal combustion
engine vehicles (ICEVs) with various alternative fuel vehicles (AVFs). Alternative fuel
vehicles are vehicles that can run on fuels other than petroleum products such as diesel or
1
Peak o il is the term used to describe the point in time at which the global conventional oil production rate
will reach its maximu m, after wh ich the annual production will begin to decline permanently [6].
2
petroleum. Alternative fuel can be electricity, compressed natural gas, hydrogen, liquid
natural gas, liquefied petroleum gas, and some biological materials like soybean or
vegetable oil-based fuels [7]. Among all alternative fuel vehicle options, the most
outstanding one is an electricity-powered one. Compared to other AFV types, electric
vehicles are more preferable regarding the fuel cost, availability of the fuel, vehicle
technology, and fuel efficiency [8-10].
Turkey has a recently growing interest in AFVs that is supported by both Turkish
government and automobile industry that are particularly focusing on electric vehicles. The
progress that is observed in the electric vehicle research and investments has already given
way to some electric vehicle models in the Turkish automobile market. Therefore, this
thesis focuses on electric vehicles rather than other types of alternative fuel vehicles in
Turkey.
There has been a great deal of studies conducted about electrical vehicles‟
advantages and importance in recent years. According to OECD 2010 reports, 23% of
world CO 2 emissions stem from the transportation sector. Thus, the decrease in carbon
emission level can be expected with the increase in use of electric vehicles in an effective
way [11]. The study of Argonne National Laboratory shows that the usage of EVs in
Houston and Washington can help to decrease CO2 emission by 26-64% and this variation
stems from the difference in electricity generation method, vehicle-recharging time per
day, and geographical location [2]. Furthermore the work of Geyer, Koehn, and Olsen
underlies that hybrid electric vehicles release around 25% less emission gas compared to
conventional vehicles [12].Moreover, many other studies support the previous argument
regarding the EV‟s positive impact on the reduction of greenhouse gas emissions,
specifically of CO2 [13-18].
Limited crude oil reserves along with the high rate of oil consumption gave birth to
studies [19-22], that show electric vehicles are effective options to reduce fuel
consumption in transportation sector. Additionally, electric vehicles are regarded as one of
the major long-term cost saving remedies with its fuel efficiency feature against the
possible high oil prices in the future [23]. Christidis, Hernandez and Georgakaki specify
that hybrid electric vehicles have 25% higher fuel efficiency compared to their
3
conventional counterparts [24]. Moreover, some type of EVs may help overcoming the
energy security problems that could develop from the dependency on imported
conventional fuels [17]. Another point that should be mentioned is the cleaner and quieter
environment that the battery electric vehicles provide compared to conventional vehicles
[25]. In addition, electric vehicles are believed that they do not need as much maintenance
as ICEVs do [26].
Although EVs seem as potential solutions for the environmental and energy related
difficulties mentioned above, penetration of EVs to the market faces certain technical and
social barriers. Immature battery technology, high price, high battery cost, and inadequate
refueling infrastructure of EVs are main technical obstacles, while the social barriers can
be listed as the lack of public knowledge on EVs, the hardships of acceptance of new
technology. Research and developments about EV technology have been continuing all
over the world to reduce the weak aspects of EVs. On the other hand, in order to eliminate
the social barriers, governments have started regulatory policies to provide subsidies or to
lower taxes on electric vehicles. Additionally, they have started adopting various
marketing strategies to raise public awareness about EVs [27,28].
Turkey has become one of the countries that recognized the significance of electric
vehicles and accordingly initiated research activities on EVs. The Scientific and
Technological Research Council of Turkey (TUBITAK) is continuing feasibility studies
and research on EVs that started in early 2000s. TUBITAK developed Electrical Vehicle
Development Platform and accelerated the research and development, and feasibility
studies on electric vehicles by the end of the 2012 [29]. Moreover, Renault, which is one of
the global automobile companies that operates in Turkey, launched new battery electric
vehicle models to the Turkish automobile market in 2011. In addition, Nissan, another
global automobile company, has been planning to introduce its battery electric vehicle
models to the Turkish market in near future. Moreover, Toyota and Honda, automobile
companies, have been offering their hybrid electric vehicle models to Turkish customers
since 2000s.
Turkish government made an important step towards eliminating the social barriers
by reducing the tax on battery electric vehicle. Furthermore, Istanbul Metropolitan
4
Municipality (IBB) supports the construction of recharging points that is required for the
infrastructure. There are already functioning charging stations located at some busy
districts of Istanbul, such as in Çamlıca, İçerenköy, Bostancı, Kartal, Florya, Avcılar and
Şişli. These developments are encouraging for the future of EV technology use in Turkey.
As a result, the aforementioned studies show that EVs appear as an important
available solution for environmental and energy related concerns. However, the
technological and social obstacles come with the EV technology cannot be disregarded.
Therefore, research and developments about EV technology, and new EV-specific policy
regulations should be carried out which would likely accelerate the adoption of EV by the
potential customers. In this regard, penetration process of EVs in Istanbul will be analyzed
considering all advantages and obstacles in the study.
1.1. History of Electric Vehicles
In recent years, due to environmental and energy related concerns, electrical vehicles
are proposed for replacing conventional vehicles. However, they are not new technologies.
Emergence of electric vehicles (EVs) goes back to as early as 1800s. First experimental
lightweight electric cars were used in the USA, the UK and the Netherlands in the mid1830s. With the developments in electro chemistry, Belgian Gaston Planté invented the
first lead-acid battery cell that is still used in most electric vehicles and in all internal
combustion engines. First electric vehicle was demonstrated with the lead battery of Planté
in France by Gustave Trouvé in 1881. During the same era, in the USA and the UK, other
similar electric cars were also developed. In 1901, Thomas Edison invented the nickel- iron
battery that could store 40% more energy per weight compared to lead battery. However,
the nickel- iron battery had a higher cost of production. Then, higher quality batteries such
as nickel-zinc and zinc-air were invented. The period from 1880 to 1900 was the golden
age of electric vehicles because most of the technological developments, which form the
basis of modern electric vehicle technologies, took place in this period. In those two
decades, new transportation options were explored with the intention of shifting from horse
drawn carriage to the more advanced options; electric vehicles, steam engine cars, and
internal combustion engine cars [25]. Although electric vehicles were effectively used at
the beginning, after a while sales of vehicles with internal combustion engines (ICE)
5
started to dominate the market particularly because of the rapid advancement of ICE cars,
limitations of batteries, and the higher cost of electric vehicles [30].
In the golden age, new battery types, and the basic principle of hybrid cars were
developed. It is worth mentioning that one of the inventors of the first hybrid vehicles was
Ferdinand Porsche. Thousands of hybrid and battery electric vehicles were produced in the
early 1900s. However, the concept of hybrid car did not become popular at the time mainly
due to the high production cost of hybrid cars during the World War I up until 1990s.
Around 1960s, there was a common debate on environmental pollution that the
conventional cars were discussed as one of the main reasons [25]. In addition, price of oil
increased rapidly in 1970s due to the Arab oil embargo. During that unlucky period of
ICEs, electrical vehicles seemed as an appropriate solution to avoid the mentioned
problems. Thus, number of experiments and research on electric vehicles were conducted
in 1970s [30]. Moreover, in the early 1990s, hybrid once again became one of the
important topics of policy discussions on cars. Some car manufacturers like Honda and
Toyota launched their hybrid car models to the market during 1990s. These models
succeeded to survive in the ICEV dominated market. Innovations and research and
developments (R&D) about cost and performance of electric vehicle, and recharging
infrastructure have been continuing since 1970s [31]. Although ICEV still dominates the
automobile market and EV still needs to be improved to meet the current demands of
customers, the diffusion of EVs has gradually but surely started all around the world with
their recognition as a significant potential solution for the environmental and energy
related problems.
1.2. Types of Electric Vehicles
Electric vehicles (EV) can be categorized under three major groups which are battery
electric vehicles (BEV), hybrid electric vehicles (HEV), and plug- in hybrid electric
vehicles (PHEV).
An electric vehicle that does not have an internal combustion engine and uses
portable battery as the only energy source is called as a battery electric vehicle (BEV).
BEVs use an electric motor instead of an internal combustion engine for the traction. BEVs
6
are electricity-powered vehicles and they can be recharged externally. BEVs have zero
tailpipe emission, high efficiency, and they operates silently [32,33].
A hybrid electric vehicle is a vehicle that has both an internal combustion engine and
an electric motor/generator [34]. Hybrid vehicles cannot be recharged externally and they
do not have portable battery. They are powered from both petroleum-based fuels and
electricity. HEV uses efficiency improvement technologies such as regenerative breaking,
and idle-off. Most of the HEVs use an electric motor as a generator that converts kinetic
energy of the moving car into the electric energy in order to charge the battery while
vehicle is decelerating. Moreover, some HEVs reduce fuel consumption by turning the ICE
off when it is idle and restart it when needed. Even this process solely could save fuel
around 5-8% [35]. HEVs are classified as series hybrid and parallel hybrid. A series hybrid
vehicle uses ICE consistently at the highest efficiency point during frequent stops and
starts. Hence, it provides lower fuel consumption in a city driving cycle. Besides, a parallel
hybrid vehicle uses ICE at the highest efficiency point while car is going at a stable speed.
Thus, it lessens the fuel consumption in the highway driving cycle [34].
PHEV is a kind of hybrid electric vehicle, which has a larger battery than a hybrid
vehicle. It can recharge its battery with electricity from off-board sources like an electric
utility grid [36]. PHEV uses two distinct powers which are chemical fuel (as HEV does)
and electricity in the battery (as BEV does) [37]. PHEV has longer driving range and lower
emission level compared to CV and HEV [38]. HEVs can be converted into PHEVs either
by replacing the existing battery pack or by adding high-energy battery pack. High-energy
battery pack stores electrical density that comes from external recharging, and from
regenerative breaking, and then stored electrical energy is send to traction motor system
[34].
HEVs could provide high fuel economy; nonetheless, they need petroleum to
operate. On the other hand, PHEVs have fuel flexibility; they can utilize either electrical
energy or petroleum depending on the demand of the vehicle driver and the level of battery
charge. On the other hand, BEVs do not use any petroleum during operation; so, they have
zero tailpipe emission, but the generation process of electricity may cause releasing CO2
7
depending on the energy source. Furthermore, BEVs rely on immature battery technologies
that are long charging times, high amount of battery cost, and limited driving ranges [36].
Comparative properties of CV, BEV, HEV, and PHEV are briefly given in Table 1.1.
Table 1.1. Comparative properties of vehicle types.
CV
Internal combustion engine
BEV
√
Electric motor
√
Portable battery
√
Usage of petroleum based fuel
√
Usage of electricity
√
External charging
√
Zero tailpipe emission
√
HEV
PHEV
√
√
√
√
√
√
√
√
√
√
8
2. PROBLEM DEFINITION AND RESEARCH OBJECTIVES
Electric vehicles have potential advantages related to environment and energy
compared to their major counterparts CVs. HEVs and PHEVs have tailpipe emission less
than CVs‟. Moreover, BEVs do not cause any emission during their operation
[19].
However, high amounts of emissions may be released during electricity generation.
Amount of emissions arisen from electricity generation varies depending on the resource
type used in generation and conversion technology. Therefore, in order to understand the
impact of electric vehicle transportation on the emission rates and global environmental
change, the emission arising both from energy generating process and tailpipe should be
taken into consideration [2]. However, extent of CO 2 emission reduction changes
depending on the demand for electric vehicles. In other words, the number of conventional
vehicles that are replaced with electric vehicles and the rate of this change at a certain
period of time are substantial when the total impact of EVs on environment is to be
assessed [39].
Potential advantages of electric vehicles support the suggestion that proposes
increasing penetration rate of EVs to the automobile market, which is dominated by the
internal combustion engine vehicles [15]. However, compared to conventional vehicles,
electric vehicles seem less likely to have the chance of having a remarkable share in the
automobile market even in the long run. Inadequate infrastructure, immature battery
technology, inadequate performance, higher purchase price, and lack of knowledge of
people about EVs could be listed as the major obstacles. Nevertheless, to which extent
market penetration of electric vehicles can be sustained is an important issue worth
mentioning. As it is stated in the history of EVs, they were firstly introduced to the public
at around 1800s, however they failed to penetrate to automobile market on account of the
reasons that were not examined in detail after the failure. Therefore, for a successful
penetration of electric vehicles into a market, it is important to understand how adoption
process could be affected by influential factors [19].
According to Rogers [40] “diffusion is the process during which an innovation is
communicated through certain channels over time among the members of a social system”
9
(p.35). Electrical vehicle penetration is included in the field of diffusion of innovation.
Diffusion of innovation models about EVs are developed with the aim of gaining insights
about how electric vehicles can initiate growth period in an automobile market [41].
However, the problem here is the diffusion of electric vehicles is hard to analyze due to
uncertainties that this new technology brings about. These uncertainties, which diffusion
normally includes in different degrees [40], can be technological, or behavioral that affect
innovation diffusion firmly. Furthermore, there are uncertain interactions among factors
[42]. Uncertainties about electric vehicle may be directly related to the electric vehicle
itself, such as future value of cost, performance, and range; indirectly related to EVs, such
as the future cost of gasoline, subsidies, and consumer preferences. Additionally, there
exists positive and negative feedbacks including various aspects such as social exposure,
infrastructure, development, technological progress in the dynamics of alternative fuel
vehicle diffusion [43].
If influential variables on EV penetration and their interactions with each other are
examined in detail considering all uncertainties, this analysis may likely help to understand
possible diffusion patterns and extents of the EV diffusion deeply. In addition, examining
EV diffusion process may help to understand how variables and interactions among them
are effective on dynamic behavior of diffusion. This assessment may help to discover
ineffective factors or the factors that may inhibit or enhance the diffusion of EVs. After
specifying these factors, inhibitor ones may be eliminated, or their affects can be
diminished with appropriate policies to support EV diffusion. Conversely, beneficial
factors may be strengthened to accelerate the diffusion process.
Diffusion of electric vehicles in a country is a comprehensive and big project that
includes the government, public and private sector, universities, and drivers. These projects
are elementarily about infrastructure, investment planning, R&D, and marketing projects.
Analyzing and understanding the influential factors on the dynamic behavior of diffusion
of EVs help to develop robust and efficient policies for diffusion.
It is important to point out that the dynamic patterns of EV diffusion process vary
from region to region particularly because of the drastic changes in population,
transportation needs, and customer preferences. Feasibility studies, infrastructure planning,
10
investment planning projects are usually developed regarding a pilot city. Therefore,
choosing an appropriate pilot city is significant to provide a better analysis of adoption. To
illustrate, in Turkey, Istanbul is the most outstanding city among options for adoption due
to two main reasons. Firstly, Istanbul is the most crowded city in Turkey. Number of
private and public car is greater than other cities. This means that transportation in Istanbul
causes more greenhouse gas emission compared to other cities. Secondly, Istanbul is the
central city of automobile market in Turkey due to broad customer profile, being close to
manufacturing plant, and high customer number. Thus, it is more reasonable to start
planning projects on Istanbul since the decline in gas emission by replacing ICE vehicles
with EVs is hand in hand with the number of potential customers that Istanbul could
provide this newly emerging sector.
One of the main objectives of this study is to answer the broad question of what are
the plausible diffusion patterns of electric vehicles for Istanbul under different scenarios
developed considering both local and global socio-economic, governmental, technological
factors and their interaction with each other? (Research question 1 (RQ1)).
In order to understand this broad question, it should be divided into three directly
related sub-questions. Firstly, this study focuses on the question of which socio-economic,
governmental, technological variables and interactions are influential on the electrical
vehicle diffusion? (RQ 1.1) Answering this sub-question helps us to discover both the
variables that have considerable influence on electrical vehicle diffusion and the variables
that show low effectiveness on EV diffusion. Secondly, the study tries to answer the subquestion that to what extent and in which direction do socio-economic, governmental,
technological variables and feedback loops affect the electric car diffusion? (RQ 1.2) This
sub-question aims to provide a deeper insight about in which direction effective variables
influence EV diffusion. In other words, answering this question assists us to comprehend
which variables and interactions among them produce what kind of dynamic behavior
patterns. In addition, impact level of these variables can be evaluated. Understanding the
variable - dynamic behavior pattern relationship makes it possible to determine variables
that induce undesirable dynamics, which can be either eliminated or reduced accordingly.
In addition, answering the second question provides the understanding that could help
strengthening and advancing the variables, which has a positive impact on diffusion.
11
Addressing RQ 1.1 and RQ 1.2 provides a better insight and understanding about
diffusion process of EVs. Thus, this situation would make it possible to answer the third
sub question that what is the expected extent of the diffusion rate for Istanbul after 3
decades? (RQ 1.3). This sub-question aims to estimate the approximate diffusion rate
under the different scenarios until 2042 considering the answers of RQ 1.1 and R.Q.1.2 and
the current transportation system of Istanbul.
Second major objective of the study is to answer the question of what are the
possible policies, which may be recommended in order to increase diffusion rate of electric
vehicles? (RQ 2). After observing the RQ1, new policies can be suggested to accelerate
diffusion process and adoption of EVs. Answering this question may help authorities such
as government, or automobile industry to understand how they should behave, or what
measures they should put into action with the purpose of quickening adoption process.
It should be noted that, a dynamic simulation model will be used in order to answer
the research questions in the model. The modeling approach and the reasons for choosing
this method will be explained in detail in the following sections. However, before the
methodology part, a number of academic works in the field of AFV diffusion that study
with a similar methodological approach will be presented in the literature section. These
works are analyzed mainly with the aim of getting a better and deeper insight about the EV
penetration topic. After the literature section, we will discuss the methodological choice
and introduce the chosen modeling approach. Afterwards, the model will be described to
put structure across completely. Subsequent to description, the model will be simulated
under different scenarios and policies. Finally, conclusion of the study will be presented.
12
3. LITERATURE REVIEW
There have been studies in search of understanding and analyzing the possible
transitions from the internal combustion engine vehicles (ICEV) to alternative fuel vehicles
(AFV). As stated earlier, a dynamic simulation model will be used to analyze EV diffusion
and to answer the research questions. For this reason, among all academic works on AFV
diffusion, the ones that use particularly simulation modeling approach are presented here to
provide better insight about EV penetration and our work. There are two different
methodologies that use dynamic simulation models to analyze EV diffusion in the
literature. They are system dynamics methodology and agent-based methodology. Firstly,
studies that use the system dynamics methodology and after works that use the agent-based
methodology will be mentioned. It should be noted that, we will highlight discussions and
outcomes that are relevant to this thesis instead of mentioning all results of the studies in
the literature section.
Struben and Sterman state that diffusion of AFVs is a dynamic, complex, as well as a
difficult process because the scale of automobile industry is too large, this large system
involves great amount of feedbacks, and it has uncertainties. For these reasons, Struben
and Sterman develop a system dynamic model to analyze possible penetration of AFVs in
the USA automobile market. This study emphasizes that especially word of mouth of nonAFVs‟ drivers are very influential on diffusion of AFVs. Word of mouth can be explained
as all activities done by drivers that may help the public to recognize AFVs such as talking
about EVs, or driving them on the road. Struben and Sterman also draw attention to
importance of sufficient recharging point for drivers as well as rapid repurchasing period to
accelerate AFV diffusion. Finally, they indicate that subsidies and marketing programs
should exist for a long period for self-sustaining adoption of AFVs [43].
Shepherd, Bonsall, and Harrison extended the dynamic model of Struben and
Sterman [43] and modified it to the UK automobile market. According to the results of
their study, 160 mile range of BEV or adequate number of recharging points would likely
suffice for BEV diffusion to be self-sustaining [39].
13
Wansart and Schnieder analyze how long-term penetration of AFV can be provided
considering its competition with the dominant conventional vehicle technology. This
model mainly focuses on effect of driving range, cost, infrastructure, and customer
awareness on AFV diffusion. They indicate that driving range capacity may limit the
maximum market share. However, this limitation does not stop the adoption process
completely. The study also draw attention that even though high battery cost and
insufficient infrastructure seem as the major obstacles of AFV penetration, low level of
customer awareness is also one of the biggest challenge that influences the long-term
adoption of BEVs [19].
According to Tran, rapid diffusion of AFVs is an important strategy in order to
decrease oil dependency and to decrease carbon emissions in the UK. For this reason, Tran
developed the model that examines both technological and behavioral sides of consumer
adoption of AFV technology. The results of the study stress that performance of AFVs
should not only catch, but also exceed the performance of ICE in order to provide rapid
diffusion of AFVs. Apart from technology, price gap between the AFV and ICE is also
considered to be one of the influential factors on diffusion in this work. Finally, importance
of supplementary polices for a rapid adoption of AFVs is emphasized [42].
Kwon utilizes a system dynamics model in order to understand market barriers of
AFVs. After understanding the barriers, he suggests policy options to eliminate them. The
major difference of this study from the previous ones is that Kwon focuses on strategic
niche management (SNM), which is a policy option to enhance market share of new
technologies such as AFVs [17]. SNM includes creating niche considering user practices,
consumers, and regulatory structures in particular areas. If niches are developed
appropriately, it offers better insights to authorities about broader application of related
sustainable technology [44]. According to the results of this study, strategic niche
management (SNM) alone may not be adequate to increase market share of AFVs.
However, it may be helpful as a reinforcing policy for financial incentives [17].
Hereinbefore, apart from system dynamics methodology, there have been studies that
use agent based methodology (ABM) to comprehend and analyze diffusion process of
AFVs. ABM considers interactions between agents since one agent can affect other agent‟s
14
decision. Multi-agent models are constructed for evaluating interactions among defined
agents that can be exemplified as consumers, automobile manufacturers, and policy makers
in the field of EV diffusion [45].
T. Zhang, Gensler and Garcia examined the factors that can accelerate the diffusion
rate of AFVs by using agent based methodology. This study mainly focuses on technology,
word-of mouth and governmental regulation. It is indicated that technological
developments of AFVs, particularly progresses about charging capacity, would likely
cause AFV diffusion to speed up. The study also supports the common observation of [43]
and [19] that is customer awareness is a strong tool to provide AFV penetration. Finally, a
high car price stands out as one of the major reasons of EV‟s small sales volume.
Therefore, some financial incentives are suggested to the government [45].
Shafiei, Thorkelsson, Asgeirsson, Davidsdottir and Raberto also developed an agentbased model in order to examine consumer behaviors and market share development of
passenger EVs. The study is carried out in Iceland and it covers ICEs and EVs. They imply
that if operating costs of PHEVs or BEVs decline, or operating costs of CVs increase, this
situation is sufficient to exceed the threshold point of diffusion process. The results of the
study continue with the argument that even though subsidy incentives may not be adequate
alone to increase diffusion rate, they can be used as supplementary policies. They also
claim that if tax on EVs is lowered, EV prices are reduced, and consumers concerns about
EVs are eliminated, penetration of EVs considerably increases. In addition to this
conditions, if gasoline prices go up, this may cause total penetration of EVs to the market.
On the other hand, reverse of these conditions come true, EV penetration becomes notably
slower. Finally, recharging infrastructure stands out as supplementary factor that may
accelerate EV diffusion [15].
Eppstein, Grover, Marshall and Rizzo also constructed an agent-based model to
study market diffusion of particularly plug- in hybrid vehicles. According to their results,
purchase price appears as a substantial factor for many consumers due to the financial
concerns. Another point that they want to draw attention is that even subsidy programs
have a positive influence on market penetration of PHEVs, its positive effect would not be
long lasting unless consumers are comfortable with PHEV technology [46].
15
In the light of these studies, certain variables that should be involved in the model for
a better analysis of the EV diffusion are roughly specified. According to the studies,
technological properties, particularly battery capacity, prices of vehicles, refueling
infrastructures of the vehicles, customer awareness as well as financial incentives are the
main aspects that the model should contain.
It is important to point out that, although at first glance, academic articles, which are
mentioned above, may seem similar to the each other as well as to our study, they differ in
many ways. Firstly, what the areas/cities where models are constructed rely on differs in
every study. In addition, structure of the model and the model boundary of every study are
different from one another. In these academic studies, most of the parameters, assumptions,
relations between variables, and formulations used in the models are not appropriate for the
automobile market in Istanbul. Moreover, Turkish customer profile, income level,
expectations, and priorities of customers, available alternative fuel vehicles types, and
regulations of the government in Turkey are substantially different from the USA and other
European countries where the earlier studies about the topic can be found. Furthermore,
examined AFV types vary in every study. Our study emphasizes particularly on the electric
vehicles while some other studies may stress on AFVs including fuel cell vehicle or
different type of EVs like plug- in hybrid vehicles. As a result, although our study and these
works have common ground; our study differs in most of ways than others.
16
4. METHODOLOGY
System dynamics is employed in this work. The reasons can be explained that the
diffusion of electric vehicles is a dynamic, non- linear, large scale and complex problem.
Variables, which characterize the diffusion of electric vehicles, change over time.
Variables interact with each other, and dynamics of electric vehicles include positive and
negative feedback loops. In this regard, system dynamics appears as an appropriate tool for
the aim of understanding diffusion dynamics of EVs through the simulation in this study.
According to Sterman [47] “system dynamics is a perspective and set of conceptual
tools that enable us to understand the structure and dynamics of complex systems” (p.vii).
In this respect, system dynamics is an effective approach in understanding and analyzing
behaviors of complex and dynamic systems. Variables, which constitute the system, are
reciprocally related. Structure of a system can be defined as the total of relationship that
exists between the variables. Causal relations between the variables creates positive as well
as negative feedback loops. These relations and loops can be visualized by causal loop
diagrams. To illustrate causal loop diagrams, a „+‟ sign on the head of an arrow means that
there is a positive causal relationship between the variable on the tail and the variable on
the head of the arrow. Positive causal relationship means that an increase (decrease) in
cause lead effect to either increase (decrease) or decrease (increase) less than what it would
otherwise have been. Conversely, a „-‟ sign on the head of an arrow means that there is a
negative causal relationship between the variable on the tail and the variable on the head of
the arrow. Negative causal relationship means that an increase (decrease) in cause leads
effect to decrease (increase) or increase (decrease) less than what it would otherwise have
been [48]. Besides, the algebraic product of all signs around the loop determines sign or
polarity of the loop. If result of the algebraic product is +, then the loop is reinforcing (R)
feedback loop. If the result is –, then the loop is balancing (B) feedback loop. An example
for causal loop diagram is given in Figure 4.1.
17
+
+ Births
R
Population
+
Birth fraction
-
B
+
+
Deaths
Death fraction
Figure 4.1. Causal loop diagram of population model.
Figure 4.1 shows that when Births goes up, Population either increases more or
decreases less than what it would have otherwise been. Besides, when Deaths goes up,
Population either decreases more or increases less than what it would have otherwise been.
Dynamics of the variables are closely related with the operation of the internal
structure. System dynamics particularly deal with the dynamics caused by internal
feedback structure of the system. Modeling, which is a scientific tool, is used with the aim
of investigating systems, problems, and solutions in system dynamics. A model is defined
as a representation of a real system with respect to a clearly stated problem. In system
dynamics models, there are two basic variables: stocks and flows. Stocks represent the
result of the accumulation over time and flows can be defined as the rate of change in these
stocks [48]. Stocks are changed by their inflows and outflows. Some examples of stocks
can be given as inventory in a manufacturing company, or number of people employed in a
firm. If stock is inventory in a company, for example, inflow and outflow may become the
production rate and the shipment rate, respectively. Another example is, if stock is number
of people employed in a firm, inflow and outflow of the stock may be hiring rate and rate
of quits, respectively [47]. In addition to stocks and flows, the third type of variable is also
used in system dynamics. It is called as auxiliary variable, or converter. Auxiliary variables
or converters help to define parameters or variables explicitly. Hence, they can be either
constant or the function of stock or the flows. An example stock- flow diagram of a simple
population model is given in Figure 4.2 [49].
18
Death fraction
Birth fraction
Birth rate
Population
Death rate
Figure 4.2. Stock- flow diagram of population model.
In the model in Figure 4.2., Population is a stock variable. Birth_ rate is the inflow of
the model. In addition, it is drained by Death rate that is the outflow of the model. Birth
fraction and Death fraction are auxiliary variables.
Population(t) = Population (t - dt) + (Birth rate - Death rate) dt
(4.1)
Birth rate = Population Birth fraction
(4.2)
Death rate = Population Death fraction
(4.3)
19
5. DESCRIPTION OF THE MODEL
The dynamic simulation model is constructed to gain a better insight about EV
penetration as well as to analyze the diffusion pro cess comprehensively. The model is
constructed regarding three types of vehicles, which are conventional vehicle (CV), battery
electric vehicle (BEV), and hybrid electric vehicle (HEV). In addition to HEVs, and BEVs,
the third kind of EV, plug- in hybrid vehicles (PHEVs), are popular in the world automobile
market. Nevertheless, they have not been sold in the Turkish automobile market actively
yet. Automobile firms in Turkey do not seem to incorporate PHEV into their product mix
in the near future. Both automobile companies and Turkish government tend to give more
importance to BEVs and HEVs compared to PHEVs. Apart from that, including PHEVs in
the model would less likely enrich the model since technical properties of PHEVs are in
between BEVs and HEVs. As a result, PHEVs are not included in the model. Apart from
EVs, gasoline-powered vehicles and diesel-powered vehicles may be categorized as
different kinds of conventional vehicles in terms of the fuel type, purchase price, and
operating cost in real life. In the model, the term of conventional vehicles (CV) is devoted
both to gasoline-powered vehicles and to diesel-powered vehicles. The parameters of CV
are composite average values that are provided from the actual data of both gasoline
vehicles and diesel vehicles. Thus, the parameters of CV represent the both type of
vehicles.
The model boundary includes only middle-size passenger vehicle market in Istanbul
(lightweight trucks, compact cars, land vehicles, buses, minibuses are excluded). So, the
parameters of vehicles, which are the driving range, refueling time, purchase price cost,
tax, operating cost, emission rates, are specified considering this particular market. It is
possible to determine various market segments and customer profiles relying on their
income level, gender, or interests in the middle-size passenger vehicle market. However,
the model is designed considering two major customer types. They are people/families
with middle income and fleet leasing companies. People/families with middle income buy
a car with the private usage aim and fleet leasing companies buy a car with the aim of
renting car to the other companies, or organization. There are three reasons to choose these
two market segments. Firstly, adding all customer profiles to the model would make the
20
model too complex that would less likely give robust results. Secondly, these two customer
profiles cover the majority of the market. Thirdly, available electric vehicle types in
Istanbul automobile market mostly fit these customer profiles. It is important to say that,
values of the related parameters are estimated regarding these customers.
The whole model is divided into six sectors that are vehicles fleet, vehicle market,
customer perception, customer awareness, infrastructure, and environmental impact in
order to describe the model eloquently. Relationships between sectors are roughly
illustrated in Figure 5.1.
Infrastructure
attributes of
vehicles
marketing
Customer
perception
Customer
awareness
non-EV drivers
Vehicle market
Vehicles fleet
Environmental impact
emission level of
one vehicle
Figure 5.1. Relations between sectors.
These sectors will be explained in detail in this chapter. However, before description,
the model will be overviewed to understand causal relations and feedback loops between
variables. The simplified causal loop diagram of the model is given in Figure 5.2.
21
+
Total CO2
+
emission
Total number of
Total demand for CV
Total demand for
EVs
Total
replenishment
Technological
development
R3
R5
CV
-
+
+
+ +
Sales of CV
+
Sales of EVs Market growth
+
Preferability of
EVs
Percentage of
unaware people
+
Total number of
EVs
+
WoM of EV
drivers
R4
+
Potential EV
customer
+
B1
+
Total social
+
exposure
+
Necessery number
of stations
R2
+
B2
+ Awareness gain
rate
+
R1
+
+
+
WoM of non-EV
drivers
Attractiveness of
vehicle
vehicle attributes
Waiting time for
refueling
-
Marketing
+
Number of
recharging stations
Figure 5.2. Simplified causal loop diagram.
As is seen in the Figure 5.2, the simplified causal loop diagram has five reinforcing
feedback loops (R) and two balancing loops (B). It should be noted that although BEV and
HEV are separately included in the model; they are represented as EV in the simplified
causal loop diagram to provide clear visualization that leads to a better understanding. The
model will be examined in detail throughout the next chapter, but, briefly, general
overviews of the main feedback loops are as follows;
R1 feedback loop represents the effect of word-of- mouth (WoM) of EV drivers on
the total number of EV in Istanbul. Word of mouth can shortly be defined as all actions
that may likely help the public to recognize EVs. For example, driving EVs on the road
and talking about EVs are involved in the word of mouth effect. Increase in the total
number of EV naturally means that increase in the number of EV drivers. That causes the
public to be exposed to increased amount of word of mouth than what it would otherwise
be. Hence, unaware people gain information about EVs and take EVs into their choice set.
Therefore, potential EV customers and demand for EV increase, and that results in an
22
increase in sales. Finally, the total number of EV in Istanbul is positively affected by sales
(vice versa).
R2 feedback loop represents word-of- mouth (WoM) effect of non- EV drivers on
potential EV customers. WoM level of non-EV drivers increases parallel with the increase
in potential EV customers. Since, as stated before, the more potential customers cause the
more people to learn about EVs even they do not drive EVs. Further, non-EV drivers
spread their information about EVs that contribute to the total social exposure. Finally, this
social exposure influences potential EV customers positively (vica versa).
R3 feedback loop represents technological developments‟ effect on sales o f EV.
Technological developments mean both research and development (R&D) as well as
learning by doing in the causal loop diagram. Raise of EV sales causes the researchers to
do more research and developments (R&D) due to two reasons. Firstly, demand in the
market may motivate the manufacturers to gain higher market share. Secondly, more
money may be allocated for R&D due to revenues coming with sales. When it comes to
learning by doing, the manufacturers find opportunity to do more practicing on production
of EVs due to higher production level. This situation leads the purchase price of EVs to
decrease. Both technological improvements and decreased purchase price cause
attractiveness of vehicles to increase and ultimately result in increase of EV sales (vice
versa).
R4 feedback loop represents the effect of recharging station number on the total EV
number in Istanbul. When the total number of EV in Istanbul increases, current recharging
stations may less likely meet the electricity recharging demands. Therefore, new charging
stations need to be constructed. New constructions causes the waiting time in a queue for
refueling to reduce that increase attractiveness of vehicle. Vehicle begin more preferable
due to rise of the attractiveness level. As a result, firstly demand for vehicle and than the
total number of EV increase (vice versa).
R5 feedback loop represents the effect of EV demand on CV demand. If demand for
CV increases, this induces CV demand to decrease. That situation leads EV demands to
increase (vice versa).
23
B1 feedback loop represents the effect of potential EV customers on awareness gain
rate. Awareness gain rate means the total number of people who recognize EVs per year.
As stated before, when potential EV customers increase, percentage of people who remains
unaware about EVs decreases. When number of unaware people decreases, awareness gain
rate normally decreases that cause potential EV customers to increase less than what it
would have otherwise been (vice versa).
B2 feedback loop represents the effect of total number of CV on EV demand via
social exposure. When total number of CV increases, WoM of non-EV drivers increases.
As stated earlier, increase of WoM level of non-EV users causes the more social exposure
to people that result in the more potential EV customers. In this situation, EV demand goes
up that result in CV demand to decline and this ultimately causes total number of CV to
decrease or increase less than what it would have otherwise been (vice versa).
Each sector will be explained in detail in the following section. Firstly, relevant
background information about sector will be given. Then, description of the related model
structure respectively will be discussed. Finally, parameter estimation and assumptions will
be presented. It is worth mentioning that only important variables, parameters, and
formulations will be given in this chapter, however all formulation and parameters can be
found in Appendix A and Appendix B.
5.1. Vehicles Fleet Sector
Vehicles fleet sector covers the variables that are sales, total number of vehicles, and
discard rates in Istanbul for all vehicle types as well as relationships of these three
variables. Vehicle fleet sector provides extensive explanation of links and formulations
between sales, total number of vehicles, and discards for each vehicle type.
5.1.1. Background Information
Number of the each vehicles type in a region firmly relies on two factors that are
sales volume and discard rate. Firstly, sales volume is shaped by customer choices about
which vehicle type they would purchase. In this model, it is assumed that consumers make
24
a choice when they purchase a car for the first time in their life, or when they replace their
current car with another one. On the other hand, discard rate, other major influential factor
on vehicle fleet, covers both vehicles that are retired, and vehicles that are sold out of the
city. In summary, formulation of the number of vehicles for each type in Istanbul is built
up on the grounds that it is filled by sales and drained by vehicles being discarded.
5.1.2. Description of the Structure
Main stock-flow structure of the vehicle fleet sector is given in Figure 5.3. This
relation is determined for each type of vehicle separately in the model.
Sales market
share of i
Total re-purchase
discard period
Total number of i
Sales of i
Discards of i
Market growth
Figure 5.3. Simplified stock- flow diagram of vehicle fleet sector.
i denotes the vehicle type. i = 1, 2, 3 mean conventional vehicles (CV), battery
electric vehicles (BEV), and hybrid electric vehicles (HEV), respectively.
Formulation of total number of i is given below (Equation 5.1)
Total number of i (t) = Total number of i (t - dt) + ( Sales of i - Discards of i) dt (5.1)
Total number of i represents total number of i-type passenger vehicles within the
boundaries of Istanbul. The total number of each vehicle type in Istanbul is a dynamic
variable. Its inflow is sales and outflow is discarded vehicles. Formulations of Sales of i
and Discards of i are shown by Equation 5.2 and 5.5, respectively.
Sales of i = Sales market share of i (Market growth + Total re - purchases)
(5.2)
25
Sales of i, which is the inflow of Total number of i, is formulated based on two
aspects that are sales market share of vehicles, and total new vehicle purchases. Sales
market share of i means percentage of total sales volume captured by i-type vehicles. This
variable will be explained in detail in Section 5.2. The second variable, total new vehicle
purchases is equal to sum of market growth and total re-purchases. Market growth is an
annual increase in the demand for a vehicle type. In other words, if person buy a car in
Istanbul for the first time in her/his life, then she/he contributes to the market growth.
Formulation of the market growth is given in Equation 5.3. Lastly, Total re-purchases, is
estimated with the sum of each vehicle type‟s discards. Vehicles are discarded when they
are broken or when they become useless due to completing their useful lifetime. Besides,
vehicles that are sold out of the city are counted as discarded vehicles in the study. Once
vehicle is discarded, customer begins to need repurchasing. Thus, discard rates of vehicles
would likely have the direct contribution to the repurchases and it is assumed that sum of
discarded vehicles produces total repurchases. As a result, sales of each type of vehicles
are estimated by multiplication of total new vehicle purchase with its own sales market
share.
Market growth = (Total vehicle demand - Total number of vehicle)/(estimation time) (5.3)
As mentioned above, Market growth is defined as an annual increase in the demand
for a vehicle. It is estimated by comparing the current demand for a vehicle in Istanbul and
the number of vehicles in Istanbul observed in the preceding year.
Total vehicle demand = Motorization rate Population in Istanbul
(5.4)
Total vehicle demand is calculated by using both Population in Istanbul and
Motorization rate of the population. Motorization rate refers to a number of passenger cars
per inhabitant. Thus, the total vehicle demand is formulated based upon the assumption
that multiplication of population and motorization rate produces total vehicle demand.
Discards of i = Total number of i / discard period
(5.5)
26
Discards of i, which is the outflow of Total number of i, captures number of i-type
vehicles that are thrown away due to being broken down, or completing useful time, and a
number of i-type vehicles, which are sent out of the city. The number of discards coming
from each vehicle type is obtained by using total number of i-type vehicle and average
discard period.
5.1.3. Parameter Estimation and Assumptions
Total number of i is a stock variable that is initialized based on the actual data taken
from Turkish Statistical Institute [50]. Motorization rate is assumed as a constant value and
it is calculated according to the current ratio of number of passenger vehicles and
inhabitant in Istanbul. In this regard, it is obtained as 0.145 vehicle /person. Population in
Istanbul changes over time and its three decades future is estimated based on population
projection of Turkish Statistical Institute [51]. It is given in Figure 5.4.
population in Istanbul
22 M
person
20 M
18 M
16 M
14 M
2012
2016
2020
2024
2028
Time (year)
2032
Population
in Istanbul
run_HEV projection of
Figure
5.4.: base
Population
2036
2040
Istanbul.
5.2. Vehicle Market Sector
Vehicle market sector consists of the sales market share of each vehicle type. This
sector provides information and formulations for each vehicle type share in the annual
sales.
27
A consumer makes a decision on a car that she/he would want to purchase among all
automobile alternatives. Her/his decision may likely be influenced both by attributes of
vehicles such as purchase price, operating cost, or size of vehicle and by personal interests.
Besides, customer should be aware of any type of vehicle to take it into her/his choice set.
Because of this, the customer‟s awareness along with the customer decisions shape the
market share of each vehicle types.
Customers are assumed to make a multi-criteria decision during purchasing process.
These decisions shape the market share of vehicles. As stated earlier, two costumer groups
in the model are people/families with middle income and fleet leasing companies. Based
on customer groups, two different market segments are determined. First market segment
is called as segment A that includes vehicles that are bought by middle-income
people/families in Turkey. In addition, the second market segment is called as segment B
that includes vehicles that are bought by fleet leasing companies. Priorities and importance
level of vehicle attributes differ from one market segment to another because customer
profiles of each segment are different. To illustrate, a fleet leasing company may have
more budget per car compared to budget of the middle- income people. Thus, purchase
price of a vehicle may have less importance for fleet leasing companies compared to the
middle- income people. Within this context, firstly, market shares of vehicles in every
segment and then overall market share of each vehicle types in Istanbul are formulated.
The sales market share of vehicle in each segment is estimated by using logit decision
model given in the work of McFadden [52]. This logit decision model is one of the discrete
choice models that are based on probabilistic consumer theory. This model is particularly
chosen because it is regarded as the most suitable one for vehicle market. Certain of
academic works in the field of vehicle market utilize from this formula [15,43].
j denotes group types (j = 1, 2 mean segment A, segment B respectively).
28
Percentege of group j
j
(5.6)
j
δi captures the sales market share of i-type vehicle among all customers. δi is
determined by using sales market share of i-type vehicle in each segment and size of
associated market segment (size means the number of vehicle belongs to associated market
segment). Percentage of group j refers to the share of related market segment in the whole
middle-size passenger car market. Formula of δi is:
δi,j captures the sales market share of i-type vehicle in market segment j. Its formula
is:
Percentage of potential customers for i
(5.7)
In Equation 5.7, ui,j refers to the total perceived utility of i-type vehicle by users in
group j. ui,j is estimated based on four utility components that are time utility, purchase
price utility, operating cost utility, and emission utility. Percentage of potential customers
for i captures the customers who are aware of i and who have i in their choice set.
Although both utility and percentage of potential customers are briefly defined in here,
they will be further described in depth in Section 5.3. and 5.4, respectively.
Percentage of the market segment A and the market segment B are estimated to be
88%, and 12% respectively based on the real data [53].
5.3. Custome r Pe rception Sector
Customer perception sector provides comprehensive description about relation
between vehicle attributes and their value perceived by customer. So, this sector consists of
attributes of vehicles and their utilities perceived by drivers.
29
While making purchasing decision about a vehicle, consumers consider all attributes
of the possible car alternatives. They compare certain attributes of vehicles with other cars
and make decision about which one they would purchase based on the total benefit
vehicles offer. Perceived utility of a vehicle represents the total benefit that vehicle offers
from the viewpoint of customers. Utility of a vehicle is firmly related to the vehicle
attributes such as driving range, refueling time, and infrastructure [15,19,36]. Moreover,
purchase price of a vehicle is one of the crucial criteria that customers care while buying a
car [43,45]. In addition, emission rate and operating costs are also effective factors on
consumer‟s preference [54]. Apart from that, two aspects are particularly determinative on
customers about accepting or rejecting new technology. They are perceived usefulness and
perceived ease of use. According to Davis [55], perceived usefulness is „the degree to
which a person believes that using a particular system would enhance his or her job
performance‟(p.320). In addition, Davis [55] says that perceived ease of use refers to „the
degree to which a person believes that using a particular system would be free of
effort‟(p.320). For example, driving range, refueling time, maintenance cost can be listed
under the perceived usefulness aspect. In addition, recharging infrastructure may be
included in the perceived ease of use aspect. In this context, it is assumed that consumers
regard six criterions during vehicle purchases: purchase price, operating cost, driving
range, refueling time, infrastructure availability, and emission rate. In addition to these
criterions, people may consider other attributes like the color, brand, acceleration, or
luxuriousness of the vehicle when making decisions. However, subjective aspects are
ignored.
Importance level of each feature may likely vary in relation to the viewpoint of
customers. For example, purchase price criteria may have more priority than emission level
criteria for most people. For this reason, when calculating utilities, evaluating each
attributes at the same importance level may result in inaccurate results. Therefore, different
importance level, which is defined as weights, is assigned to the each attribute in the
model. Apart from that, importance weights fluctuate depending on customer types since
the significance as well as the priority of attributes differ among people. For instance,
operating cost has more priority for the group of people with middle income compared to
30
the fleet leasing companies. Because the middle-income people pay for fuel from their own
budget while fuel cost is paid by customers of fleet leasing companies. Correspondingly,
same attribute have different weight in each market segment in the study. In the real world,
importance level may differentiate for every individual. However, constructing a model
considering each individual‟s priority is almost impossible because of the following
reasons. Firstly, it is unattainable to detect desire of every individual. Secondly, even it is
possible to figuring out every person‟s interest, putting every variable in the model makes
the model extremely hard, in fact impossible, to analyze. Finally, even the model includes
too many importance levels that can be analyzed; result of such kind of study may less
likely be robust. Accordingly, we specified representative customer profile for each market
segment. They capture mainly population behavior in the related segment and we set the
importance levels of attributes regarding these representative profiles.
Important parts of customer perception sector and relations between these parts are
given in Figure 5.5. This relation is determined for each type of vehicle separately in the
model.
i perceived
utility
time utility
of i
purchase price
utility of i
operating cost
utiliy of i
emission utility
of i
Effect of
infrastructure
purchase price
maintenance
refueling time
refueling cost
weight of time
cost utility
weight of
of i
utility
utility
purchase
utility
driving range
of i
weight of refueling
power-source
cost utility
unit cost
maintenance
cost
weight of
emission utility
i emission rate
weight of
maintenance cost
utility
Figure 5.5. Simplified diagram of perceived utility sector.
Consumers choose a vehicle by comparing the utilities they perceive about each
vehicle types. Four major utilities are included in the model. They can be listed as time
31
utility, purchase price utility, operation cost utility, and lastly emission utility. In all utility
types, two major factors, which are vehicle attributes and weight of these attributes, help to
estimate the utility. It is important to point out that the larger absolute value of a weight
means a higher importance level. Weights used in the study are estimated considering the
revealed-preference multinomial logit model of alternative fuel vehicle preferences
estimated by Brownstone, Bunch, and Train [56].
wikj refer to weight of utility k for i-type of vehicle for users in market segment j (k
= 1,2,3,4,5 time utility, purchase price utility, refueling cost utility, and maintenance cost
utility, emission utility respectively) Values of all weights are given in Appendix B.
5.3.2.1. Time Utility: Time utility is related to driving range and refueling time features of
vehicle, as well as availability of refueling stations on account of being forceful factors on
consumer decision. For instance, having less driving range leads to run out of fuel quickly.
In this situation, drivers have to find refueling stations, and refuel their vehicle frequently
that can be seen as a waste of time from the view of drivers, particularly for busy ones.
Besides, if finding refueling stations is not easy, drivers would have to wait for roadside
assistance or tow their vehicle to refueling stations. Even if drivers easily find refueling
stations, there may be a queue due to lack of adequate recharging points. Moreover, long
refueling time causes time loss. The less duration of refueling causes less waste of time. In
summary, less driving range, inadequate infrastructure, or long refueling time may be
costly in terms of time. Therefore, consumers regard driving range, refueling time, and
refueling infrastructure while purchasing a car. In this respect, time utility formula is
developed considering average time loss stemming from driving range, refueling time, and
refueling infrastructure in a certain distance.
Time utility formula:
tu ij
Maximum range
Refueling time of vehicle i
Driving range of vehicle i
effect of recharging infrastructure wi1 j
tuij refers to the time utility of i type of vehicle for users in market segment j
(5.8)
32
wi1j refers to weight of time utility for i- type of vehicle for users in market segment j.
wi1j helps to determine quantitatively importance level of time.
Maximum range refers to maximum range that customers desire to drive without a
need of recharging process. It is a function of daily driving habits. Driving range of vehicle
is defined as the average range of km before vehicle needs a refuel. This corresponds to
how much a conventional car and a hybrid car can drive with one tank gasoline (50 lttank), and the total range of km, which a BEV can drive with one full battery. Refueling
time of vehicle is the duration of refueling of gasoline tank/battery fully. Driving range and
refueling time are not constant and these two parameters change due to learning curve
effects for BEV. Learning curve effects provide a mean to count improvements about
battery technology in the model since improvements about battery technology would likely
continue gradually due to cumulative research and development studies. Formulation of
learning curve effect is given in Equation 5.9.
(t )
(0)
( E (t ) / E (0))
(5.9)
denotes value of the attribute at time t
denotes value of cumulative experience at time t
denotes learning factor
Learning curve effects for battery technology show accumulation of suppliers
experience on vehicle potentially turning into improvement in terms of research and
developments [57]. In the study, learning curve effect is regarded for only battery
technology. In other words, HEV technology or CV technology is not improved by
learning curve effects in the study.
Effect of infrastructure on i refueling time shows impact of refueling infrastructure
on time utility. It will be analyzed in detail in Section 5.5.
33
5.3.2.2. Purchase Price Utility:
As stated earlier, purchase price is a substantially
influential factor in the course of choosing a car. Because of this, purchase price criterion
is included in the model and its weight is relatively high. In addition to price, costumer
budget is also included in the purchase price utility formula because natural logarithm of
the budget is used to normalize purchase price of vehicles. In this context, budget is
determined as annual income level of household for middle-income people and total
money earned from one car for fleet leasing companies. In addition to budget,
normalization factor (nf) is also used to normalize purchase price. Purchase price utility
formula is given as;
pui1
Purchase price of vehicle i / nf
wi 21
ln(annual income level of household)
(5.10)
Purchase price of vehicle i/ nf
wi 22
ln(money earned from one car)
(5.11)
pui 2
puij refers to purchase price utility of i-type of vehicle for users in market segment j
wi21 refers to weight of purchase price utility for i-type of vehicle for users in market
segment A.
wi22 refers to weight of purchase price utility for i-type of vehicle for users in market
segment B.
wi2j helps to determine importance level of purchase price for market segment j
quantitatively.
Purchase price of vehicle i
i price before taxes (1 i PCT) (1 VAT) (5.12)
Purchase price of vehicle consists of price before taxes, private consumption taxes,
and value added taxes. Private consumption taxes (PCT) and value added taxes (VAT),
which are determined by the government, are added to the price before taxes. That
produces purchase price of a vehicle. Another important point here is that even purchase
prices of vehicles may be relatively high currently; prices would decline gradually due to
34
learning curve effect that comes from learning-by-doing of manufacturers. Learning by
doing helps manufacturers to gain experience about production of vehicle. Experiences
ultimately would yield a decrease in production cost that induces decline in purchase price
costs. Therefore, it is assumed that purchase prices of BEV, HEV, and CV will decrease
gradually over time.
5.3.2.3. Operating Cost Utility: Operating cost is constituted of fuel cost and maintenance
cost in the model. Thus, operating cost utility is equal to sum of refueling cost utility and
maintenance cost utility.
ouij
(5.13)
ru ij mu ij
ouij refers to operating cost utility of i-type of vehicle for users in market segment j
ruij refers to refueling cost utility of i-type of vehicle for users in market segment j
muij refers to maintenance cost utility of i type of vehicle for users in
market segment j
ruij
refueling cost
i
wi3 j
(5.14)
wi3j refers to weight of refueling cost utility for i-type of vehicle for users in market
segment j. wi3j helps to quantitatively determine importance level of refueling cost for j
market segment.
Refueling cost is defined as the cost of fuel that vehicle uses in order to travel one
km distance. Fuel types, which are supplied from refueling stations, are gasoline for
conventional and hybrid cars, and electricity for battery electric vehicle.
mu ij
maintenanc e cost
i
wi 4 j
(5.15)
35
wi4j refers to weight of maintenance utility for i-type of for users in market segment
j. wi4j helps to quantitatively determine importance level of maintenance cost for users in
market segment j.
Maintenance cost is defined as normalized version of monthly maintenance cost of
vehicle. Monthly maintenance cost consists of normalized versions of both battery renting
cost (if portable battery is available) and routine monthly maintenance cost. BEVs need
portable battery, which can be purchased or rented to operate. However, purchase price of
a battery is extremely high, thus drivers mostly prefer to rent the battery. Therefore, it is
assumed that battery is rented monthly. Routine monthly maintenance cost refers to cost of
maintenance that every vehicle should have in every month. In the maintenance cost
calculation, BEV maintenance cost is taken as a base value for normalization. Maintenance
cost of both CV and HEV are normalized based on BEV maintenance cost.
5.3.2.4. Emission Utility: Emission is defined as total CO 2 released from conventional
and hybrid vehicles during their trip. However, there is no tailpipe emission coming from
battery electric vehicles. Therefore, emission of BEV is specified as CO 2 that is released
to the atmosphere from electricity plant during electricity generation. Moreover, emission
level is total CO 2 level released for a km drive of CV, HEV, and BEV.
Formulation of emission utility is given in Equation 5.16.
euij
emission rate
i
wi5 j
(5.16)
euij refers to emission utility of i type of vehicle which belongs to group j
wi5j refers to weight of emission utility for i-type of vehicle which belongs to group
j. wi5j helps to determine importance level of emission for j market segment quantitatively.
Emission utility of vehicle is equal to multiplication of vehicle-normalized version of
emission rate and weight of emission utility. For normalization, emission rate of
conventional vehicle is chosen as a base value. Accordingly, normalized version of
emission level of HEV and BEV are estimated.
36
Choosing appropriate attribute values for each car type in the model appears as a
challenge due to the broad profile of the middle-size passenger vehicles. In order to
overcome this challenge, representative vehicle profiles that mainly capture market
segment of the middle- income people and fleeting companies are determined for CV,
HEV, and BEV. After that, value of each attribute is estimated by taking the average value
of vehicles that are fit to that profile.
Available battery technology currently provides 165 km range [58] and 1-hour
refueling time [59] for BEVs. Maximum range is determined heuristically considering the
area of Istanbul and it is assumed as 300 km. However, sensitivity tests are applied to the
maximum range.
There are fundamentally three types of recharging stations based on their power and
refueling time in the EV market. Those are home-type, normal, and quick charging
stations. Their charging time is about 6 hours (usable at night at home, or parking area), 36 hours (usable in parking area of shopping malls), 0.5-2 hours (recharging stations)
respectively [59]. Long charging is commonly used when vehicle is idle whereas quick
charging is mostly used during trips if vehicle runs out of the fuel. In the model, time
utility of a vehicle that is during operation is estimated. Thus, only quick recharging
stations are considered in the model. Refueling time for BEV is assumed to be an hour on
average. It is assumed that recharging points are distributed evenly across Istanbul.
Moreover, recharging infrastructure affects time utility via effect function. Effect
function of infrastructure is given in Figure 5.6.
37
Figure 5.6. Effect function of infrastructure on BEV time utility.
effect of recharging infra. on tu2 j
f(
BEV number per station
) (5.17)
referenceBEV number per station
Effect functions can help to show non- linear relations between variables. In this
context, effect function, which is seen in Figure 5.6, is used to show non- linear relation
between time utility and recharging infrastructure. Reference BEV number per stations is
called as normal value. If BEV number per station exceeds normal value, then time utility
begins to be affected negatively from this situation. For example, f(1) represent the
situation that BEV number per station is equal to normal value. Figure 5.6 shows that if
BEV number per station becomes 12.5 times more than reference BEV number, then time
utility of BEV becomes one fourth of f(1).
Annual income level of a household is determined regarding middle- income families.
For this reason, household income is determined as 5000 TL per month that equals
annually 60000 TL. Besides, money earned from one car for fleet leasing companies is
determined as 80000 TL. It should be noted that these two values are used for
normalization of the purchase price. Natural logarithms of these values are included in
purchase price utility formula.
The government has offered a new regulatory law about private consumption tax
(PCT) for BEVs for incentive. PCT for vehicles that have only electric motor is determined
as 3%, 7%, and 15% depending on motor power in accordance with the new law. It is
38
chosen as 3% due to the motor power of representative BEV. However, this value is 37%
for both CVs and HEVs. Value added tax is 18% for all vehicles types [60].
Refueling cost is estimated regarding current prices of gasoline and electricity.
According to [61] gasoline vehicles consume about 8 lt/100 km. Current price of gasoline
is about 4.8 TL/lt in Turkey. Thus, cost per km is estimated to be 0.384 TL. This value is
25% less for hybrid cars due to efficiency of HEV. Besides, BEV uses 0.2 kWh per km on
average [61,62]. Electricity cost for households is 0.0563 TL/ 0.2 kWh (0.0563 TL/ km) on
average. However, the model includes only commercial refueling and in this sense,
electricity cost is determined considering a profit margin.
Tailpipe CO2 emissions of conventional vehicle is about 188 g/km [64]. Emission
rate of a hybrid car is about 25% less than a conventional car [12]. As it is stated earlier,
although BEVs release zero tailpipe emission, certain amount of CO 2 is created during the
electricity generation process. Amount of CO 2 depends on energy sources. Emission level
stemming from BEV varies among countries because of the using different energy sources.
For example, in Turkey, around 362.8 tons CO 2 is released to the atmosphere for one GWh
electrical energy production [65]. As it is said before, BEV uses 0.2 kWh per km on
average [61,62]. Therefore, emission arisen from electricity generation process for one
BEV to drive one km is estimated to be 72.56 g/km in Turkey based on actual data taken
from Turkish Electricity Transmission Company [65].
5.4. Custome rs Awareness Sector
Customer awareness sector includes social exposure coming from marketing and
drivers as well as impact of social exposure on customer familiarity with a vehicle. This
sector provides explanation for relation between social exposure and customer awareness.
Any vehicle type can enter choice set of consumer, if and only if consumer is aware
of that vehicle type. Therefore, awareness of people about vehicles is a substantial factor
39
for purchasing decisions [19,43]. As a result, customer awareness is analyzed in depth in
this work.
Everybody in Istanbul is aware of conventional vehicles. On the other hand, EVs are
new technology and Turkish customers are not completely familiar with the EV concept. If
customers gain awareness of the EVs, they become potential EV customers and take EVs
their choice set during purchasing a car. However, being potential EV customer requires
sufficient social exposure because drivers need cognitive and emotional process to
consider EVs during purchasing vehicle. Social exposure also accelerates this emotional
process. It is assumed that social exposure, which helps people to become potential EV
customers, is provided by marketing activities and word-of- mouth (WoM) of people about
EVs in this work. Marketing activities cover all marketing channels such as TV
advertisements, newspapers, journals, magazines, and internet. Moreover, word-of- mouth
includes all ways that drivers can spread information about EVs on their own suc h as
conversation, driving EV on the road, internet, or social media.
Important parts of customer awareness sector and relations between these parts are
given in Figure 5.7.
EV awareness loss
fraction
Percentage of non-EV
+ drivers who aware of EV
Percentage of
potential customers
for AFV
EV customers
awareness loss
marketing influence
on EV
EV Customer
awareness gain
Total social +
exposure +
Effectiveness of word of
mouth of non-EV drivers
Social exposure of
non-EV drivers
Social exposure of
+ Percentage of EV
+ EV drivers
Effectiveness of word of
drivers
mouth of EV drivers
Figure 5.7. Simplified stock- flow diagram of customer awareness sector.
40
New types of products, particularly innovative ones, necessitate consumer- learning
process for adoption. EVs are also innovative products that need process to be accepted by
customers. Percentage of potential EV customers is a term that is used to represent portion
of customers who accept EV technology emotionally and cognitively in the model.
Percentage of potential EV customers refers to a percentage of customers who are willing
to take EVs into their consideration set during purchasing. It is assumed that customers are
equally familiar with HEV and BEV, potential EV customers have all information and
understanding about EVs. If customers gain awareness about EVs, they become potential
EV customers.
Percentage of potential EV customers (t)
Percentage of potential EV customers (t - dt)
(Customer awareness gain
(5.18)
- Customer awareness loss) dt
Percentage of potential EV customers is formulated using familiarity model of
Struben and Sterman [43]. Potential EV customers is filled by awareness gain and drained
by awareness loss.
Customer awareness gain
Total social exposure
(5.19)
(1 - Percentage of potential EV customers)
Customer awareness gain is shaped by multiplication of total social exposure and
percentage of customer, who are not aware of EVs. In other words, unfamiliar people with
EVs learn about EVs through social exposure.
Total social exposure Marketing influence Social exposure of EV drivers
(5.20)
Social exposure of non - EV drivers
Total social exposure refers to all social influences that help people to recognize
EVs, and learn information about EVs. Total social exposure is equal to the sum of
marketing, social exposure of EV drivers, and social exposure coming from non- EV
drivers. Marketing is an important strategy in launching process of new products. It helps
peoples to recognize product or it shows profitable sides of product. Apart from marketing,
41
both social exposure of adopters and non-adopters of the technology spread via word of
mouth (WoM) of drivers. Word-of- mouth includes all activities that drivers can spread
information about EVs such as conversation, driving EV on the road, internet, or social
media. These activities make people gain information about electrical vehicles and take
EVs into their consideration set. Formulations of social exposure of EV drivers and nonEV drivers are given in Equation 5.21 and 5.22, respectively.
Marketing influence is the effect of marketing on percentage of potential EV
customers.
Social exposure of EV drivers (Effectiveness of WoM of EV drivers)
V EV
Vt (5.21)
VEV denotes the total number of EV (sum of BEV and HEV) in Istanbul.
Vt denotes the total number of vehicle in Istanbul.
Social exposure of EV drivers is equal to multiplication of fraction of total EVs
driven in Istanbul and effectiveness of WoM of EV drivers. Number of adopters is
substantially important factor for adoption because when people talk to vehicle owners or
see them on the road, they recognize the availability of EVs and learn about them. Thus,
the more adopter would likely cause more aware people. Effectiveness of WoM of EV
drivers represents quantitative estimation of potential effectiveness of EV-drivers on
consumer awareness level.
Social exposure of non - EV drivers (Effectiveness of WoM of non - EV drivers)
Percentage of potential EV customers
(1 -
(5.22)
VEV
)
Vt
Social exposure of non-EV drivers stems from drivers who are aware of EVs but not
owing one. It is shaped both by fraction of the non- EV number whose drivers are aware of
EV driven in Istanbul and by effectiveness of WoM of non-EV drivers. Effectiveness of
42
WoM of non-EV drivers represents quantitative estimation of effectiveness of non-EVdrivers consumer awareness level.
Customer awareness loss Percentage of potential EV customers
EV awareness loss fraction
(5.23)
Customer awareness loss is equal to multiplication of percentage of potential
customers for EVs and EV awareness loss fraction. Awareness loss is included to the
model because even people learn about EVs; some of them may take EVs out of their
consideration set after a while unless their perception is reinforced by continuing social
exposure. This situation causes potential EV customers to decrease. EV awareness loss
fraction represents percentage of potential consumers that forget about EVs.
It is assumed that every driver in Istanbul is aware of CVs and CV is included in all
drivers‟ choice set. Therefore, percentage of potential CV customers is stable variable and
it is set to 100% throughout simulation.
Effectiveness of WoM of EV drivers and Effectiveness of WoM of non-EV drivers is
determined to be 0.25 and 0.15, respectively. EV drivers are more influential compared to
non-EV drivers because they directly experience EV technology. This situation makes
them more persuasive. In real life, marketing influence is a dynamic variable and it
depends on budget coming from sales. However, in the model, marketing influence is
assumed to be a constant value. It is important mentioning that these values are estimated
considering the study of Struben and Sterman [43]. Sensitivity tests are also applied to
these values.
EV awareness loss fraction is equal to 1% per year. In other words, every year, 1% of
potential EV customers forgets about EVs and takes EV out of their consideration set.
Sensitivity tests are also applied to the EV awareness loss fraction.
43
5.5. Infrastructure Sector
Infrastructure sector provides explanation about refueling infrastructure of vehicles.
This sector covers all infrastructural process of vehicles.
Refueling infrastructure sufficiency means that number of refueling stations
adequately meets the refueling demand of all vehicles. Refueling infrastructure sufficiency
is a serious factor for driver comfort because inadequate infrastructure may cause drivers
to wait for roadside assistance or tow their vehicle to refueling stations in the case of
running out of the fuel. Besides, there may be a queue on the recharging points due to the
lack of adequate recharging stations. These circumstances would likely result in time lost
during trip, which may be very important for a driver. As a result, infrastructure
sufficiency is an important criterion for customers [15,19] and this factor is incorporated to
the model.
BEVs differ from CVs and HEVs in terms of refueling points. CVs are gasolinepowered vehicles that use gasoline stations for refueling. BEVs are electricity-powered
vehicles that use electricity-recharging points. When it comes to HEV, it uses gasoline
stations because it does not supply electricity from external source. The point is that
gasoline stations are currently adequate for CV and HEV in Istanbul. Thus, refueling is not
a problem for both CV and HEV drivers. However, inadequate infrastructure is one of the
substantial concerns about BEVs [54]. There is potential and growing demand for BEVs
and current number of recharging stations wo uld not meet this potential demand.
Construction of recharging point for BEV continues and total station number is a dynamic
variable. Hence, it is examined in detail whereas number of gasoline stations for CV and
HEV is assumed as stable and sufficient for vehicles.
Recharging stations can be built under two different strategies; proactive or reactive.
In proactive strategy, stations are constructed before diffusion of BEVs in order to prepare
infrastructure of the city to BEV usage. Thus, it is expected that there would be no obstacle
arisen from infrastructure in the proactive strategy. On the other hand, stations are
44
constructed right after the penetration of BEVs in the reactive strategy. Current number of
BEVs in the city would be estimated and necessary station number would be determined.
Recharging points would be built considering this estimation under reactive strategies.
Station construction policy taken by Turkey is exactly same neither proactive nor reactive
ones. However, in the base model, reactive strategy is regarded because it seems that
recharging points construction would begin to speed afterwards increase of BEVs sales in
Turkey. Therefore, current infrastructure policy, which is applied in Turkey, resembles
closer to reactive strategy.
Important parts of infrastructure sector and relations between these parts are given in
Figure 5.8.
Desired station per
vehicle ratio
Current number of
stations
Gap
Desire number of
BEV stations
Number of
Recharging Stations
of BEVs
construction
Number of BEV stations
planned to be constructed
Municipality
criteria
planned
construction
Planning delay
Effect of desired
constraction on nmb of
recharging stations
Desired
constraction
Construction delay
Figure 5.8. Simplified stock- flow diagram of infrastructure sector.
As stated before, gasoline stations are currently adequate for CVs and HEVs
Istanbul. Thus, refueling infrastructure is not a problem for CV and HEV drivers. On the
other hand, there are currently thirteen charging points in Istanbul and this number appears
45
to be notably inadequate value when potential BEV demand is considered. However, new
constructions would be implemented in parallel with beginning of EV penetration. In this
context, number of recharging stations is formulated and it is given in Equation 5.24.
Number of recharging stations of BEVs (t)
Number of recharging stations of BEVs(t - dt)
construction dt
(5.24)
Number of recharging stations is a stock variable and it is changed by construction.
Construction is defined as a number of recharging points being completed per year.
New charging points are built in the case of current ones begin insufficient to cover the
total demand. The recharging point number that shows sufficient value to supply electricity
to all BEV market is defined with the term of desired number of BEV stations.
Formulation of this term is shown in Equation 5.25.
Desired number of BEV stations Perceived number of BEV in Istanbul
Desired station per vehicle ratio
(5.25)
Desired number of BEV station refers to a necessary station number in Istanbul for
drivers to find recharging points easily and not to wait in a queue for a long time. It is
equal to multiplication of perceived number of BEVs in Istanbul and desired station per
vehicle ratio. Perceived number of BEV in Istanbul is a smoothed version of the total
number of BEVs in Istanbul. Besides, Desired station per vehicle ratio is a ratio that is
formulated regarding how many station points per vehicle should be available in Istanbul
to sustain adequate infrastructure. In other words, after total number of BEVs is estimated
with a delay, its multiplication with desired station per vehicle ratio produces desired
number of station in Istanbul.
Lastly, the gap between desired number of stations and current ones may such a huge
that it may be hard to cover due to limited budget or feasibility studies. Therefore, it is
46
assumed that there is an upper municipality criterion, which restricts construction of
stations.
Number of Recharging Stations of BEVs is initiated with thirteen stations that is the
current number of recharging points in Istanbul. Desired station per vehicle ratio is
determined to be 0.05 station/vehicle regarding refueling time, and accessibility of stations.
In real life, the municipality, or private companies may implement construction of
recharging stations. However, there has been no clarified information about who would be
responsible for recharging points‟ construction, or what would be regulations. Thus, in the
base model, it is assumed that both municipality and private companies may construct
recharging points but there is an upper limit for number of annual construction for
recharging stations due to budget, and proper area constraints. It is assumed that maximum
1000 station per year can be constructed (municipality criterion). Besides, sensitivity
analyzes are applied to the desired number of stations per vehicle ratio and the
municipality criterion.
5.6. Environmental Impact Sector
Environmental impact sector includes CO2 reduction coming from EV penetration.
This sector provides understanding for possible CO 2 reduction associated with fleet.
There are various kinds of emission gases arisen from transportation industry such as
hydrocarbons, CO, NOx , and CO2 . In the study, when environmental impacts of EVs are
estimated, only CO 2 is regarded on account of the two reasons. Firstly, scientific studies
show that CO 2 plays the most significant role in the transportation related climate changes.
Secondly, CO 2 amount stemming from vehicles is substantially higher compared to the
other emission gases [66]. As mentioned earlier, most of researches claim that EVs may
likely be an effective solution for CO2 emissions. However, this substantially depends on
both number of EVs that replace CVs, and means of electricity generation. For these
47
reasons, once diffusion rate of EV is observed, its effect on CO 2 will be analyzed in order
to estimate ultimate environmental impact of EVs.
Impact of the EV diffusion on the environment is traced comparing two different
cases. In the first case, CO 2 emission rate is estimated assuming that all vehicles in Istanbul
are CV and thus, CO 2 can only arise from CV. In other words, it is estimated regarding that
how much CO 2 would be emitted if there were not any penetration of EVs. In the second
case, CO 2 stemming from associated fleet sizes of CV, BEV, and HEV are estimated.
Finally, these two cases are compared and impact of EV diffusion on CO 2 emission is
estimated. Formulation of CO 2 reduction in every year is given in Equation 5.26. In
addition, formulation for cumulative CO 2 reduction is given in Equation 5.27.
Reduction of CO
2
Total emission level when all cars are CV - Total emission level
(5.26)
Total emission level when all cars are CV
Cumulative CO 2 reduction (t) Cumulative CO 2 reduction(t - dt)
CO 2 emission difference dt
(5.27)
CO 2 emission difference Total emission level when all cars are CV
- Total emission level
(5.28)
Total emission level when all cars are CV Annual range of CV Emission level of CV
Total vehicle in Istanbul
(5.29)
Total emission level when all cars are CV represents amount of CO 2 if all cars on the
road are conventional vehicles. Annual range of CV means total range that a CV is driven
in a year. Emission level of CV is equal to the sum of emission levels arisen from each CV.
48
Total emission level
Emission level of total i
(5.30)
i
Total emission level represents the total amount of CO 2 coming from BEVs, HEVs,
and CVs in Istanbul. To observe total emission level, firstly, emissions arisen from each
vehicle groups are separately calculated and then estimated values are summed.
Emission level of total i Annual range of i Emission level of one i
Total number of i in Istanbul
(5.31)
Emission level of total i represents total annual amount of CO 2 released from all i type vehicles. Emission level of one i equals to amount of CO 2 released by one i-type
vehicle for a km-drive. Multiplication of emission level of one i with annual drive
produces annual CO 2 level released from one car. Then, total emission level of i is
estimated for total i-type cars. Annual range of i means total range that a vehicle is driven
in a year.
Annual range of vehicles is determined as 18000 with assuming daily travel range of
a vehicle as 50 km.
49
6. VALIDATION AND ANALYSIS OF THE MODEL
Vensim software is used for running the simulation model. It is continues time model
and integration type is Euler. In this context, for all the simulation runs, time step is
selected as 0.125. This value is neither too large to give inaccurate results nor too small to
cause computer calculation errors.
The time unit of the model is taken as a year. The time horizon of the simulation is
set to three decades, from 2012 to 2042 in order to be long enough to capture direct,
indirect, and delayed effects of the variables and feedbacks.
In this chapter, firstly, validation of the model will be discussed in Section 6.1. Right
after validation, base behavior of the model will be presented in Section 6.2.
6.1. Model Validation
Model validation is an important step of system dynamics methodology that checks
if the model is an acceptable and adequate representation of the system with respect to
dynamic problem of interest. Model validity is tested both in structural and behavioral
aspects [48]. For testing the model validity, two major test groups, structural and
behavioral, are applied to the model. These major groups and their related sub-groups will
be extensively explained in the following sections.
6.1.1. Structural Validity
Structural validity tests analyze if the structure of the model can reflect the actual
relations that exist in the real problem of interest meaningfully and satisfactorily. These
tests should be established before behavior validity tests because if the structure of the
model is invalid, then its behavior becomes unreliable. Structure validity involves two
distinct tests that are direct structure tests and structure oriented behavior tests [67].
50
Significant portion of structural validation has been done during the model construction
process. For example, variables of the model have real life counterparts. Moreover, the
structure of the model is a meaningful description of the real relations that exist in the
problem. Besides, direct structure tests and structure oriented behavior tests are applied
after construction. Some of direct and structure oriented behavior tests will be given below.
6.1.1.1. Direct Structure Tests: Direct structure tests analyze the model structure validity
by direct comparison with knowledge about real system structure. Structure and parameter
confirmation tests, dimensional consistency tests, and direct extreme conditions tests are
covered by the direct structure tests group [48]. For example, all variables and parameters
in the model have real life counterparts. To illustrate, all parameters about attributes of the
vehicle types, current vehicle number, and station numbers that are involved in the model
are estimated from empirical data. Furthermore, each equation and relationship is
compatible with available knowledge about the real system. In addition, there is
dimensional consistency in the model (Units can be found in the Appendix B). Finally, the
model equations are evaluated to be valid under extreme conditions.
6.1.1.2. Structure Oriented Behavior Tests: Structure oriented behavior tests analyze the
validity of the structure indirectly. In this regard, extreme condition tests and sensitivity
analysis, which are the two basic test groups under structure oriented behavior tests, are
performed. Firstly extreme condition tests and following that sensitivity analysis will be
clarified.
Extreme condition tests (ect): Extreme condition tests help to understand if the
model is robust under extreme conditions or not. Robustness under extreme situations
means that the model should behave in a realistic way independent from how extreme
policies are applied to the model [47]. In this context, three different extreme condition
tests are applied to the model. They are given below.
Extreme condition test 1 (ect 1): In the first extreme condition test, BEV attributes
are modified. Its refueling time is set to 20 hours, and its driving range is set to 5 km. In
addition, purchase price of vehicle is shifted from 50000 TL to 100000 TL. In these
51
conditions, it would be expected that there would be no BEV sales due to insufficient
attributes of BEV. Sales market share of EV is given in Figure 6.1.
BEV sales market share
100
percent
75
50
25
0
1
2012
1
1
1
1
1
1
1
1
BEV sales market share : extreme_condition_1
1
1
1
1
1
2016
1
1
2020
1
1
2024
2028
Time (year)
2032
2036
1
1
1
1
2040
Figure 6.1. BEV sales market share under the 1st ect.
As can be seen from the Figure 6.1, nobody buys BEV. The results that were reached
are matching with the expected outcomes because in first extreme situation, battery electric
vehicle attributes become not satisfactory for consumers due to both extremely low value
of time utility and high value of purchase price. Thus, it is logical that nobody prefers BEV
and market share becomes zero. As a result, the model is valid under extreme condition
test 1.
Extreme condition test 2 (ect 2): In the second extreme condition test, marketing
influence on customer awareness is modified. As mentioned earlier, inflow of percentage
of potential EV customers is directly affected by sum of marketing influence, social
exposures of EV drivers and non-EV drivers. Normally, marketing influence is equal to
0.01 in the base model. However, it is set to 0.99, which means that marketing exposure on
people becomes quite intense. It is expected in this test that percentage of potential
customers reaches 100% fast and cannot be greater than 100% naturally no matter
marketing influence is notably high. Behavior result of percentage of potential EV
customers is given in Figure 6.2.
52
Percentage of potential customers for EV
1
100
1
1
1
1
1
1
1
1
1
1
1
1
1
percent
75
50
25
1
0
2012
2016
2020
2024
2028
Time (year)
Percentage of potential customers for EV : extreme_condition_2
1
1
2032
1
1
2036
1
2040
1
1
1
Figure 6.2. Percentage of potential EV customers under 2th ect.
Figure 6.2 shows that when marketing influence is set to 0.99, percentage of
potential EV customers converges to one but it does not exceed it. The result is similar to
what is expected because independent from the variables, percentage of potential EV
customers cannot be greater than 100% in the nature of things. As a result, the model is
valid under extreme condition test 2.
Extreme condition test 3 (ect 1): In the third extreme condition test, attributes of the
BEV and HEV are improved equally and extremely. In other words, their driving ranges
are set to the 1000 km, refueling time, maintenance costs, operating costs and purchase
price are decreased excessively. Both emission rates are set to zero. Conversely, attributes
of the CV are deteriorated as much. Purchase price, operating cost, maintenance cost, and
refueling time are increased immensely. After this modification, behaviors of customers
who are aware of all type of vehicles are analyzed. In this test, it is mainly expected that
50% of customers who are familiar with EVs would choose BEV whereas other 50%
would choose HEV. In addition, it is expected that no one would prefer CV due to
insufficient attributes of CV. Figure 6.3, Figure 6.4, Figure 6.5 show shares in annual sales
of BEV, HEV, and CV among potential EV customers respectively.
53
BEV sales market share in potential EV customers
100
percent
75
50
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
25
0
2012
2016
2020
2024
2028
Time (year)
BEV sales market share in potential EV customers : extreme_condition_3
1
2032
1
1
2036
1
2040
1
1
1
Figure 6.3. BEV share among potential EV customers under 3rd ect.
HEV sales market share in potential EV customers
100
percent
75
50
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
25
0
2012
2016
2020
2024
2028
Time (year)
HEV sales market share in potential EV customers : extreme_condition_3
1
2032
1
1
2036
1
2040
1
1
1
Figure 6.4. HEV share among potential EV customers under 3rd ect.
54
CV share market share in potential EV customers
100
percent
75
50
25
0
1
2012
1
1
2016
1
1
2020
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2024
2028
Time (year)
CV sms in potential EV customers : extreme_condition_3
2032
2036
1
1
1
1
2040
Figure 6.5. CV share among potential EV customers under 3rd ect.
Results show that 50% of people who are familiar with EVs prefer BEVs and other
50% of people who are familiar with EVs prefer HEVs. In addition, no one prefers CVs
among potential EV customers. The results behave in expected direction because if
attributes of CVs becomes much worse than HEVs‟ or BEVs‟, nobody wants to buy CVs.
Moreover, if BEVs and HEVs exactly have same attributes, utilities coming from them
become at same level that cause their market shares to be equal. These results indicate that
the model is valid under 3rd extreme condition test.
In conclusion, the model is valid under these three extreme condition tests.
Sensitivity analysis: Sensitivity analysis is another sub-group of structure oriented
behavior tests. According to Barlas [67], „Behavior sensitivity tests consist of determining
those parameters to which the model is highly sensitive, and asking if the real system
would exhibit similar high sensitivity to the corresponding parameters‟(p.4). In the study,
most of the parameters are tested to understand whether there are parameters, to which the
model is highly sensitive or not. In this regard, sensitivity tests are applied to all
parameters that do not have exact values taken from their real counterparts. Results of all
analysis can be found in Appendix C. However, an example set, which contains
effectiveness of WoM of EV drivers, motorization rate, and weight of emission utility, will
be analyzed in detail below. Another important point is that the range of sensitivity is
55
arranged between plus 20% and minus 20% of the base value in all tests. In other words,
the minimum value of range is determined as 20% less than the base value of parameter
and maximum value of range is specified as 20% higher than the base value of same
parameter in sensitivity tests.
Effectiveness of word of mouth of EV drivers: In the model, quantitative estimation of
effectiveness of WoM of EV drivers is assumed as 0.25 (dimensionless). In the sensitivity
analyses, this value is tested between the range of 0.2 and 0.3. The impact of this
modification on the BEV fleet market share is given in the Figure 6.6.
base run
BEV fleet market share
2042
30
20
10
0
Time (year)
Figure 6.6. Sensitivity result for effectiveness of WoM of EV-drivers.
The Figure 6.6 shows that BEV fleet market share is not strongly sensitive to
effectiveness of WoM of EV-drivers.
Motorization rate: Motorization rate is assumed as 0.145 vehicle/person in the
model. Although this value is observed from current habitat motorization rate, sensitivity
tests are applied to this parameter, too. This value is changed between the range of 0.115
and 0.175 (vehicle/person). The impact of this adjustment on the BEV fleet market share is
56
base run
2042
15
10
5
0
Time (year)
Figure 6.7. Sensitivity result for motorization rate.
Figure 6.7 shows that BEV fleet share seems relatively insensitive to motorization
rate.
Weight of emission utility: Weight of emission utility is assumed to be -0.07
(dimensionless) for market segment A, and -0.09 (dimensionless) for market segment B.
These values are adjusted based upon the revealed-preference multinomial logit model
estimated by Brownstone, Bunch, and Train [56]. Nonetheless, sensitivity tests are applied
to the weights to assess their influence on the model sensitivity. In this test, weight of
emission utility for A is changed between the range of (-0.085) and (-0.056). In addition to
A, for B, the range is arranged between (-0.011) and (-0. 7). Impact of this modification on
the BEV fleet share is given in Figure 6.8.
In conclusion, all results (the results given here and results in the Appendix) indicate
that the model is valid in terms of parameters that are max range, effectiveness of WoM of
EV drivers, effectiveness of WoM of non-EV drivers, motorization rate, every estimation
time, annual range of vehicles, weight of every utility. It must be noted that there may be
sensitivity in numerical results. However, the model has low sensitivity in terms of pattern
dynamics. This means that long-term behavior of the model strongly depend on structure
of the model rather than some uncertain variables.
57
base run
2042
15
10
5
0
Time (year)
Figure 6.8. Sensitivity result for weight of emission utility.
Figure 6.8 shows that in given range, impact of weight of emission utility do not
substantially varies.
6.1.2. Behavior Validation
Once model succeeds structural tests, behavior validity tests are applied to control if
the dynamic patterns, which the model produces, are close enough to the real patterns of
concern. Behavior validation tests assess the pattern prediction, not point prediction. This
type of validation involves some statistical and quantitative tests like regression and trend
comparison, periods and amplitude comparison, or BTS software [48]. However, EVs,
particularly BEVs, have been a topical issue in Turkish automobile market for less than 2
years. Thus, Istanbul or Turkey has no historical data about battery electric vehicles or
hybrid vehicles to compare with behavior patterns of the simulation result of the model.
However, the model generates patterns that resemble the ones perceived in other mobility
systems. Behavior patterns show similar trajectories the ones in observed in the literature.
BEV and HEV fleet share results of this study and an example result from the study in the
literature are respectively given in Figure 6.9 and Figure 6.10 to show pattern similarity.
An example result is taken from the work of Wansart and Schnieder [19].
58
Fleet market shares BEV and HEV
40
percent
30
20
2
2
2
10
0
12
2012
2
1 2
12
2016
1 2
12
1
2020
BEV fleet market share : base run
HEV fleet market share : base run
2
1
2
1
2
2
1
2
1
1
1
1
1
2
1
2
1
2024
2028
Time (year)
2
1
2032
1
2
1
2
2036
1
2
1
2
2040
1
2
1
2
1
2
Figure 6.9. Fleet share patterns of this study.
Figure 6.10. Fleet share pattern of the work of Wansart and Schnieder.
6.2. Analyses of the Base Behavior
The base model is a case that forms a reference for comparison and assessment of the
scenarios and policies in the analysis. In the base case, it is assumed that the recent trends
would likely continue with no major changes. In other words, all technological
improvements, prices, costs, and regulations given in the beginning of the analysis would
59
be progressed gradually according to broadly accepted trajectories. Driving range,
refueling time, and maintenance cost of BEV improve gradually over time. Purchase price
of BEV, HEV and CV decrease gradually over time. It is assumed that both electricity and
gasoline prices would show similar trend to their historical data. Finally, emission rates are
assumed to be constant for all vehicle types.
The base model includes three types of vehicles that are conventional vehicle (CV),
battery electric vehicle (BEV), and hybrid electric vehicle (HEV). Customers buy a vehicle
among three types with comparing attributes of each vehicle. Preferences of customers
shape market share of each vehicle type. It must be noted that two different market shares
are defined in the study. They are sales market share and fleet market share. Sales market
share represents percentage of total sales volume captured by each vehicle in the market. It
is a share in the annual flow. On the other hand, Fleet market share represents the ratio of
number of each vehicle type to the total number of vehicle available on the road of
Istanbul. It is a share in the active stock. Sales, which fill amount of each vehicle type, are
shaped by sales market share. Ratio of number of each vehicle type to the total vehicle
number produces fleet market share of related vehicle type.
The behavior pattern of sales market share of vehicles is illustrated in Figure 6.11.
Sales market share
100
3
3
3
3
3
3
3
percent
75
3
3
3
50
25
0
1 2
2012
12
12
2016
1 2
12
12
2020
HEV sales market share : base run
BEV sales market share : base run
CV sales market share : base run
2024
2028
Time (year)
1
1
2
3
3
2032
1
2
1
2
3
3
3
1
2
3
12
2040
1
2
3
1 2
12
1
2
3
3
2036
1
2
3
12
1 2
1 2
12
12
1 2
3
1
2
3
3
Figure 6.11. Sales market share of vehicles under the base run.
60
As stated earlier, Sales market share represents percentage of total sales volume
captured by each vehicle in the market. As is seen from the Figure 7.1, after three decades,
sales market share of each of BEV and HEV reaches 30.64% and 30.21% respectively.
There are two main reasons why sales EV market share is still lower than CVs‟ even after
3 decades. Firstly, although percentage of potential customers goes up gradually, there are
still people who do not recognize EVs yet. These unaware customers directly buy CV due
to the perception of the unavailability of other choices. Secondly, although attributes of
BEV and HEV may have preferable sides compared to CV, they remain still less efficient
than the attributes of CV. For instance, driving range of BEVs is lower, refueling time of
BEVs is longer, or maintenance cost of BEV is higher compared to CVs‟ throughout
simulation period. In addition, purchase price of HEV is higher than both BEV and CV.
Apart from these, sales market share of HEV is slightly higher than BEV‟s sales market
share in the first 25 years of the simulation. After this point, BEV sales begin to catch and
exceed HEV sales. This means that some attributes (maintenance cost utility and time
utility) of HEV are more preferable from the view of customers until the last years of the
simulation. However, improvements about battery technology provide BEVs to be more
advantageous.
Sales of new vehicles add to the total number of each vehicle type in Istanbul, which
is given in Figure 6.12.
Total number of vehicles
3M
vehicle
2.25 M
3
3
3
3
3
3
3
3
3
3
3
3
3
1.5 M
750,000
0
12
2012
12
12
2016
12
12
2020
total number of HEV : base run
total number of BEV : base run
total number of CV : base run
12
12
12
2024
2028
Time (year)
1
1
2
3
12
1
2
3
1
2
3
1
2
3
12
12
2036
1
2
3
12
2032
1
2
3
12
3
2040
1
2
1
2
3
12
1
2
3
3
Figure 6.12. Total number of each vehicle under the base run.
61
According to the simulation results, total number of every type of vehicles increases
in the first 10 years. The reason of why total number of CV increases while sales market
share of CV decreases is the growth of automobile market in Istanbul. After the first
decade, while number of BEV and HEV continue to increase, CV starts to decline due to
the high market share captured by BEV and HEV.
Total number of each vehicle forms the fleet market share of vehicles. Fleet market
share represents the ratio of number of each vehicle type to the total number of vehicle
available on the roads of Istanbul. Fleet market share of each vehicle type is presented in
Figure 6.13.
As it could be seen from the Figure 6.13, even the sum of BEV and HEV cannot
achieve to capture 1% market share in the first years of diffusion. After 2016, penetration
gains speed and increases gradually. BEV and HEV diffusion ultimately reaches
respectively around 19.76%, and 20.7% of the total fleet in Istanbul by 2042. The reasons
of the increase in the fleet share can be explained as rise of familiarity with EV, improved
BEV technology, fall in price of BEV and HEV as well as developed BEV recharging
infrastructure. All these circumstances lead EVs to become more attractive for customers
that result in EV market share to increase over time.
Fleet market share
100
3
3
3
3
3
3
3
3
3
3
3
percent
75
3
3
3
50
25
0
1 2
2012
12
12
2016
1 2
12
2020
CV fleet market share : base run
1 2
12
12
1 2
12
2024
2028
Time (year)
1
2
1
2
3
1
2
3
2032
1
2
3
3
2036
1
2
1
2
3
3
2040
1
2
1
2
3
12
1 2
12
12
1 2
1
2
3
3
Figure 6.13. Fleet market share of vehicles under the base run.
62
As stated in the previous part, sales firmly depend on awareness about vehicle.
Percentage of potential customers for CV is equal to 100% due to the assumption of 100%
familiarity of costumers to the conventional vehicles. However, percentage of potential
customers for EV is not 100% and it is increased by marketing activities, and word of
mouth of adopters and non-adopters. In this sense, observed simulation result of potential
customers for EV is given in Figure 6.14.
Percentage of potential customers for EV
100
1
1
1
1
75
1
percent
1
1
50
1
1
25
1
1
1
0 1
2012
1
1
2016
2020
2024
2028
Time (year)
Percentage of potential customers for EV : base run
1
2032
1
2036
1
1
2040
1
1
1
Figure 6.14. Percentage of potential EV customers under the base run.
Figure 6.14 illustrates that the percentage of potential EV customers has an S-shaped
behavior pattern since it grows slowly in the beginning due to the lower number of
adopters and higher number of non-adopters. In addition, major population among nonadopters does not have knowledge about EVs. After a while, it grows faster because both
non-adopters who are familiar with EVs and adopters increase.
One of the main reasons of why EVs are suggested to replace CVs is the potential
reduction in greenhouse gas emissions. In this work, ultimate impact of EV diffusion on
CO 2 reduction is estimated. Reduction percentage is observed based on EV penetration
data coming from the simulation results. It should be noted that reduction of CO 2 does not
mean
cumulative
reduction.
It
means
Figure 6.15 shows CO 2 reduction under base run.
reduction
in
the
associated
year.
63
Reduction of CO2
20
1
15
1
percent
1
1
10
1
1
5
1
1
1
0
1
2012
1
1
2016
1
1
2020
Reduction of CO2 : base run
1
2024
2028
Time (year)
1
1
1
1
2032
1
2036
1
1
2040
1
1
1
Figure 6.15. Reduction of CO 2 under the base run.
The Figure 6.15 shows that CO 2 reduction related to EV penetration would be around
5% at 2028. In addition, even with both 19.76% fleet market share of BEV and 20.77%
fleet market share of HEV, reduction of CO 2 emission would only reach around 17% in
2042. In addition, cumulative CO 2 reduction would be around 17.07x106 tons by 2042.
As it seems in the base run figures, S-shaped pattern is observed for the sales market
share of both BEV and HEV. This is mainly due to the fact that domination of CV in the
beginning of penetration process causes low level of familiarity of consumers about EV.
Thus, market diffusion of EVs shows slow growth in the early periods despite
competitiveness of EVs in terms of operating costs, purchase price, and emission rate.
Even after familiarity level increases, market share of neither BEV nor HEV reach the
market share of CV due to the two reasons. Firstly, even if awareness level increases, there
are people who do not recognize EVs throughout simulation. Secondly, certain preferable
properties of conventional vehicles, which are lower purchase price and maintenance cost
compared to both BEV and HEV, and higher time utility compared to BEV induce higher
CV market share.
64
7. SCENARIO AND POLICY ANALYSIS
In this section, eleven different scenarios, ten policies, and four scenario-policy
combinations are evaluated to explore how they would influence diffusion pattern and
market share of EVs in Istanbul. Each scenario, policy, and combination will be explained
in detail below.
7.1. Scenario Analysis
In scenario analysis section, eleven different scenarios are examined to capture
plausible changes in the context. Topics of these scenarios are basically future costs of
electricity and gasoline, BEV technology, refueling infrastructure, launching only BEV in
to the market, customer awareness, and repurchasing rate. The results of these scenarios
will mostly be presented comparing them with the base run to provide better understanding
for analysis.
7.1.1. Electricity and Gasoline Costs Related Scenarios (Scenario 1)
Battery electric vehicles are electricity-powered vehicles whereas their conventional
and hybrid counterparts use gasoline as a power source. Because of this reason, costs of
electricity and gasoline may likely be effective factors on customer decisions about vehicle
types. However, future prices of electricity and gasoline are uncertain. Thus, it would be
better to specify possible trends about electricity and gasoline and then to analyze impact
of these trends on EV penetration.
Resources used in electrical energy generation and their usage rates differ among
countries. Despite the fact that renewable resources (such as wind power) can be used in
electricity generation, natural gas, brown coal, imported coal, hydraulic resources are basic
resources that have considerably high shares in electricity generation in Turkey. In
addition, some of these resources (such as high percentage of natural gas or some portion
of coal) are imported, which can be limited amount or have high tax or price. Electricity
price is highly related to the shares of different resources in the generation mix. Cost of the
65
energy resources is also influenced by exogenous factors like economical or political
development [68]. In brief, there are many factors on electricity price that cause
uncertainty of the future electricity price in Turkey.
Future prices of gasoline is also an uncertain factor for vehicle diffusion because
gasoline prices may be dramatically affected by global fuel prices, exchange rates, and
political and economical developments [68]. These reasons make it hard to predict future
trends of gasoline prices.
In the base case, it is assumed that both electricity and gasoline prices would show
similar trend to their historical data. Historical electricity prices data shows fluctuation.
Therefore, it is hard to determine a constant annual raise percentage for electricity price.
However, in the base case, annual increase rate of electricity price is assumed as 9% on
average. This value is an estimation based on historical trend of electricity prices. In
addition, historical data about gasoline prices show that increase rates of gasoline prices
highly fluctuate and thus, it is hard to determine annual raise percentage due to great
variance. However, in the base case, annual raise percentage is determined as constant
value that is forecasted as 7.5% on average based on historical trends.
Four scenarios about the future situation of energy prices are constructed to analyze
their effect on EV penetration. In the first scenario, it is assumed that there would be no
further changes in the value of the gasoline and electricity prices throughout simulation
period. It is assumed that electricity cost is slightly affected by BEV demand s in the second
scenario whereas electricity cost is firmly affected by BEV demands in the third scenario.
Lastly, rapid increase in gasoline prices is assessed in the final scenario. These four
scenarios will be analyzed in detail in the following part. It should be noted that the term of
cost does not exactly represent power prices. Gasoline cost and electricity cost are
estimated based on the fuel usage rate of vehicle per km and prices of resource that they
use. In other words, cost refers to operating cost of a vehicle for a km drive. However, cost
directly reflects increase or decrease of the gasoline and electricity prices.
7.1.1.1. Constant Electricity and Gasoline Costs (Scenario 1_1): In this scenario, it is
assumed that there will be no factor that can influence gasoline or electricity prices
66
directly. Therefore, gasoline and electricity costs will not change throughout simulation.
The result of Scenario 1_1 on BEV and HEV fleet shares are illustrated in Figure 7.1 and
Figure 7.2, respectively.
20
1
2
1
15
2
1 2
percent
12
1 2
10
12
1
5
0
12
2012
2
1 2
12
1 2
2016
1 2
12
2020
1 2
12
2024
2028
Time (year)
BEV fleet market share : Scenario 1_1
1
1
2
2032
1
2
1
2
2036
1
2
1
2
2040
1
2
1
2
1
2
Figure 7.1. BEV fleet market share under the Scenario 1_1.
HEV fleet market share
40
percent
30
20
10
0
12
2012
1 2
12
2016
1 2
12
2020
1 2
12
12
1 2
1 2
2024
2028
Time (year)
HEV fleet market share : Scenario 1_1
1
1
2
2032
1
2
12
1 2
1
2
2036
1
2
1
2
2040
1
2
1 2
12
1 2
1
2
1
2
Figure 7.2. HEV fleet market share under the Scenario 1_1.
As can be seen from the graphs, if gasoline and electricity costs do not change,
diffusion of BEVs would be lower compared to the base run because in the base run,
67
gasoline and electricity costs keep rising and the gap between them gradually widens. This
situation influence BEV diffusion positively. However, in this scenario, cost gap remains
the same.
7.1.1.2. Low Level of Sensitivity to Electricity Demand (Scenario 1_2): There has been a
debate about whether increase of electricity demand due to BEVs affects electricity prices
or not. In this scenario, annual increase percentage of gasoline is assumed as constant and
7.5%. However, annual rise percentage of electricity price is not constant in this case. It
positively depends on total electricity supply. In other words, the more electric vehicle
causes more electricity consumption. This situation may lead to rise of electric prices. In
this scenario, it is assumed that when BEV share in fleet reaches 5%, electricity prices
begin to be affected by electricity demand slightly. In other words, sensitivity of electricity
prices to electricity demand is at low level in this scenario. Pattern of electricity and
gasoline prices and their impact on BEV fleet market share are respectively given in Figure
7.3. and Figure 7.4.
Electricity vs gasoline costs
4
3
3
3
TL/km
3
3
2
2
3
3
1
3
3
0 12
2012
1 2
1 2
2016
1 2
3
3
3
3
12
2020
3
12
12
1 2
2
1 2
1 2
2024
2028
Time (year)
Electricity unit cost : base run
1
Electricity unit cost : Scenario 1_2 2
CV gasoline unit cost : Scenario 1_2
1
1
2
1
2
3
12
1
2
2032
1
2
3
2
3
3
3
2036
1
2
1
2
3
1
3
2040
1
2
1
2
3
1
1
1
2
3
3
Figure 7.3. Electricity and gasoline costs under the Scenario 1_2.
68
40
percent
30
20
1
1
1
10
0 12
2012
1
1 2
12
2016
1 2
12
2020
12
1 2
2
1
12
1 2
2
2
2
2
1 2
2024
2028
Time (year)
1
BEV sales market share : Scenario 1_2 2
1
2032
1
2
1
2
2036
1
2
1
2
2040
1
2
1
2
1
2
In this scenario, electricity cost increases to its maximum, 2.42 TL/km, by 2042. As
seen from the figures, if electricity cost is slightly affected by electricity usage rate, this
situation does not create considerable change on BEV sales in the first 25 years of the
diffusion. However, after 25 years, it begins to influence sales strongly because after 25
years high- level electricity demand causes a balancing mechanism against diffusion.
7.1.1.3. High Level of Sensitivity to Electricity Demand (Scenario 1_3): As stated in
Scenario 1_2, electricity prices can be influenced by electricity demand. On the other hand,
unlike previous one, in this scenario, it is assumed that the electricity prices are highly
affected by electricity demand coming from BEV drivers. After a while, electricity cost
exceeds gasoline cost due to price-demand relation. It is assumed that when BEV share in
fleet reaches 5%, electricity prices begin to be affected by electricity demand. After this
point, electricity prices continue to increase parallel in with increase of BEV share.
Moreover, it is assumed that annual rise of gasoline price would follow its historical trend,
which is 7.5% per year. Although the end of the simulation is determined to be 2042 at
whole analysis, the model is simulated until 2050 to provide visual clarity only for this
scenario.
69
The simulation results show that the gap between gasoline and electricity costs
begins to narrow at around 2040. In addition, electricity cost exceeds gasoline cost at 2045.
After 2040, BEV sales begin to decrease while HEV continue to go up. After breakeven
point is exceeded, BEV sales begin to decline increasingly.
15
TL/km
11.25
7.5
1
2
3.75
0 12
2012
12
1
2
2016
1
2
2020
1
2
2024
Electricity unit cost : Scenario 1_3
Gasoline unit cost : Scenario 1_3
2028 2032 2036
Time (year)
1
2
1
2
1
2
2040
1
2
1
1
1
1
1
1
1
2
2
2
2
2
2
2
12
1
2
2044
1
2
2048
1
2
1
1
2
2
Figure 7.5. Electricity vs gasoline prices under the Scenario 1_3.
Sales market share
45
1
1
33.75
1
percent
1
1
1
22.5
2
2
2
2
2
12
2
1 2
12
11.25
12
0 12
2012
12
2016
12
1 2
2020
2024
2028 2032 2036
Time (year)
HEV sales market share : Scenario 1_3
BEV sales market share : Scenario 1_3
1
2
1
2
1
2
2040
1
2
2044
1
2
1
2
2048
1
2
1
2
Figure 7.6. Sales market share of BEV and HEV under the Scenario 1_3.
70
Fleet market share
30
1
1
1
22.5
1
percent
1
1
15
2
2
2
2
2
12
12
7.5
12
0
12
2012
12
12
12
2016
12
12
2020
2024
2028 2032 2036
Time (year)
1
2
1
2
2040
1
2
1
2
2044
1
2
1
2
2048
1
2
1
2
7.1.1.4. High Gasoline Cost vs Normal Electricity Cost (Scenario 1_4): Gasoline prices
may rise due to the reasons such as political or economical issues, relationship of countries,
supply problem, or new tax regulations. In the last scenario, it is assumed that gasoline cost
increases normally (7.5% in every year) until 2020. After 2020, it shows exponential
increase. However, electricity cost increases normally (9% per year) throughout
simulation. Patterns of sales and fleet market share of vehicles under Scenario 1_4 are
respectively given in Figure 7.8 and 7.9.
8
TL/km
6
2
2
4
2
2
2
2
0 12
2012
1
2
2
1
2016
2
1
1
1
2020
Electricity unit cost : Scenario 1_4
Gasoline unit cost : Scenario 1_4
2024
2028
Time (year)
1
2
1
2
2032
1
2
1
2
2036
1
2
1
2
2040
1
2
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
1
2
1
2
Figure 7.8. Gasoline vs electricity cost under the Scenario 1_4.
71
Sales market share
40
2
2
2
percent
30
1
1 2
2
1
1
1
12
20
1 2
1 2
1
10
1
0 12
2012
1
1 2
1
2
2016
2
1 2
2
2
2020
2024
2028
Time (year)
1
2
2032
1
2
1
2
2036
1
2
1
2
2040
1
2
1
2
1
2
Figure 7.9. Sales market shares of BEV and HEV under the Scenario 1_4.
Fleet market share
30
22.5
percent
2
1 2
15
1
1
1 2
12
1
7.5
0
1 2
2012
12
12
2016
12
12
2020
1 2
1 2
12
2024
2028
Time (year)
1
2
1
2
2
12
2032
1
2
2036
1
2
1
2
2040
1
2
1
2
1
2
Figure 7.10. Fleet market shares of BEV and HEV under the Scenario 1_4.
Results show that the gap between electricity and gasoline costs widens gradually
until around 2030. However, the gap begins to increase rapidly after this year. Thus, after
around 2030, BEV sales exceed HEV sales and BEV sales continue to show its existing
trend. However, HEV sales firstly begin to increase decreasingly and then begin to
decrease. This shows that if gasoline cost increases substantially while electricity cost
keeps rising gradually, after a while sales market share of HEV begins to decline.
72
The summarized results of scenarios related to changes in electricity and gasoline
costs are given in Table 7.1.
Table 7.1. The results of electricity and gasoline costs related scenarios.
Market Share
2012 (%)
Market Share
2042 (%)
Total sales until 2050
(Million)
Base Run
CV
99.99
59.47
6.990
BEV
0.003
19.76
1.032
HEV
0.007
20.77
1.131
Scenario 1_1
Constant electricity and gasoline costs
CV
99.99
61.33
7.071
BEV
0.003
17.94
0.954
HEV
0.007
20.73
1.127
Scenario 1_2
Low level sensitivity to electricity demand
CV
99.99
59.76
6.999
BEV
0.003
19.19
1.014
HEV
0.007
21.05
1.139
Scenario 1_3*
High level sensitivity to electricity demand
CV
99.99
59.96
6.552
BEV
0.003
18.80
1.113
HEV
0.007
21.24
1.485
Scenario 1_4
Higher gasoline cost vs normal electricity cost
CV
99.99
55.83
6.923
BEV
0.003
22.07
1.125
HEV
0.007
20.10
1.103
*Scenario 1_3 is simulated until 2050. However, the values on the Table 7.1 are
results in 2042. Shares in fleet of CVs, BEVs, and HEVs by 2050 are 50.73%, 20.37%, and
28.90%, respectively.
7.1.2. Technological Development Related Scenarios (Scenario 2)
Vehicle attributes are influential factors on customer preferences about vehicle types.
Although BEVs have competitive properties such as lower operating cost or lower
emission rate compared to CV and HEV; BEVs may fall behind them due to the
insufficient infrastructure and the limited battery properties such as lower driving range,
longer refueling hour, or higher maintenance cost.
73
Research and developments have been continuing all over the world to improve BEV
technology because researchers imply that technological improvements about the battery
technology may yield penetration of BEV to speed up [27,42,43,45]. Although it is
believed that technological improvements about battery are going to occur, possible
improvement level and time are highly uncertain. In this sense, three possible scenarios
about technological improvement are assessed. It is assumed that progress level of battery
technology would gradually grow in the base case. After that, first scenario is developed
assuming moderate technological improvement whereas it will be at optimistic level in the
second scenario. Finally, the case of no technological improvement is analyzed in the third
scenario.
7.1.2.1. Moderate Technological Improvement (Scenario 2_1):
In this scenario,
improvement about battery technology is assumed to happen at medium level. In other
words, driving range, refueling time, and maintenance costs are improved more than the
base run, but eventually, they do not become as efficient as conventional vehicles even
after three decades. Progresses of these three properties are respectively presented in
Figure 7.11, 7.12, and 7.13. In addition, impact of these improvements on fleet shares is
Driving ranges of vehicles
1,000
3
3
3
3
3
3
3
3
3
3
km
750
4
500
250
1 2
0
2012
1
1
2
2016
2
2
2020
1
1
1
2
2
2024
2028
Time (year)
2
2
2032
1
1
1
1
1
2
2036
2
2
2040
BEV driving range : Scenario 2_1
1
1
1
1
1
1
1
BEV driving range : base run
2
2
2
2
2
2
2
HEV driving range : base run
3
3
3
3
3
3
3
CV driving range : base run 4
4
4
4
4
4
4
4
Figure 7.11. BEV driving range under the Scenario 2_1.
74
Refueling times of vehicles
1
1 2
12
2
2
1
0.75
hour
2
1
2
1
2
1
2
1
2
1
2
1
2
1
1
0.5
0.25
0
2012
3 4
34
3 4
2016
34
2020
3 4
34
3 4
2024
2028
Time (year)
34
34
2032
34
2036
3
2040
BEV refueling time : Scenario 2_1
1
1
1
1
1
1
1
BEV refueling time : base run
2
2
2
2
2
2
2
HEV refueling time : base run
3
3
3
3
3
3
3
CV refueling time : base run 4
4
4
4
4
4
4
4
Figure 7.12. BEV refueling time under the Scenario 2_1.
Maintenance costs of vehicles
1
2
1
1
1
1
1
1
1
Dmnl
0.75
1
1
1
0.5
3
0.25
4
0
2012
2016
2020
2024
2028
Time (year)
BEV maintenance cost : Scenario 2_1
BEV maintenance cost : base run
HEV maintenance cost : base run 3
CV maintenance cost : base run
4
1
2
1
2
3
1
2
3
4
2040
1
2
3
4
2036
1
2
3
4
2032
3
4
1
2
3
4
4
Figure 7.13. BEV maintenance cost under the Scenario 2_1.
75
Fleet market share
25
2
2
18.75
percent
2
1
2
12.5
1
2
1
6.25
1
1
1
2
1 2
12
0
1 2
2012
1 2
12
2016
12
12
2020
12
2024
2028
Time (year)
1
2
1
2
2032
1
2
2036
1
2
1
2
2040
1
2
1
2
1
2
Figure 7.14. Fleet market share of BEV and HEV under the Scenario 2_1
As is seen from the Figure 7.14, HEV has slightly higher market share than BEV in
the first years, but after a while, improvements of the battery technology cause BEV to
exceed the market share of HEV. Improvements do not create substantial change on the
fleet share because despite of the medium- level developments, driving range, refueling
time, and maintenance cost are still less preferable than CV and HEV.
7.1.2.2. Optimistic Improvements (Scenario 2_2): In this scenario, it is assumed that BEV
technology would immensely be improved and some of attrib utes of BEV would be better
than CV technology over time. Moreover, there would not be any infrastructural obstacles
for BEVs throughout the simulation. In other words, the coupled effects of evolving
technological features and advanced infrastructure on adoption process are captured in this
scenario. It should be noted that developments of battery attributes are parallel with
optimistic scenarios given in the literature. Advanced progress of technological features
and infrastructure situation is illustrated in Figure 7.15. Moreover, effect of this scenario
on BEV fleet share is given in Figure 7.16.
76
Figure 7.15. Technological and infrastructural improvement under the Scenario 2_2.
30
22.5
2
1
percent
2
15
2
2
2
7.5
2
2
0
1 2
2012
1
2
12
12
2016
12
1
2020
2
1
2
1
1
1
1
1
2024
2028
Time (year)
1
1
1
1
2
1
2
2032
1
2
2036
1
2
1
2
2040
1
2
1
2
1
2
Figure 7.16. BEV fleet market share under the the Scenario 2_2.
The results show that as it is expected, advanced technological improvements and
sufficient recharging stations cause BEV diffusion to increase. However, even if all
technological and infrastructural conditions are advanced like in the Figure 7.15, fleet
market share of BEV would achieve to reach 23.46% of the market by 2042. This is mainly
77
due to the fact that the driving range of BEVs exceeds driving range of CVs after 25 years,
refueling time and maintenance cost of BEVs are improved but they still does not keep
pace with or exceed CVs‟ attributes.
7.1.2.3. No Improvement (Scenario 2_3): In the last technological development related
scenario, it is assumed that there would be no improvement about technology, or reduction
of the purchase price of BEVs. In addition, it is assumed that construction of recharging
points to be continued. However they remain insufficient to meet the recharging demand
throughout the simulation. This scenario is the worst-case scenario about BEVs. Results of
this scenario are given by comparing results with the base run and Scenario 2_2.
The results of the last scenario indicate that even if there were no improvements
about technology, purchase price, and infrastructure, BEV may succeed to penetrate 9.52%
of the market. Furthermore, Figure 7.18 shows that if BEV technology remains at its
current level, more customers will choose HEVs and CVs compared to the base run. This
situation causes HEVs and CVs to capture higher market share.
30
percent
22.5
2
1
2
2
15
2
2
2
7.5
0
1 23
2012
1
2
2 1
12 3 12 3 1 3
2016
2020
2 31
2
2
3
1
1
2024
2028
Time (year)
1
1
2
1
2
3
2032
1
2
3
3
2036
1
2
1
2
3
3
2040
1
2
1
2
3
3
3
3
3
3
3
3
3
1
1
1
1
1
1
2
3
3
78
30
3
22.5
3
percent
3
1 2
3
15
1 2
3
12
3 1
2
7.5
1
1 2
3 12
3 12
0
1 23
2012
12 31 2
12 3 12 3
2016
31 2
2020
31 2
2024
2028
Time (year)
1
1
2
1
2
3
2032
1
2
3
1
2
3
2036
1
2
3
1
2
3
2040
1
2
3
1
2
3
3
Figure 7.18. HEV fleet market share under the Scenario 2_3.
The summarized results of scenarios related to changes in technology are given in
Table 7.2.
Table 7.2. The results of technological development related scenarios.
Market Share
2012 (%)
Base Run
CV
BEV
HEV
Scenario 2_1
CV
BEV
HEV
Scenario 2_2
CV
BEV
HEV
Scenario 2_3
CV
BEV
HEV
99.99
0.003
0.007
Market Share
2042 (%)
(Million)
59.47
6.987
19.76
1.036
20.77
1.129
Moderate technological improvements
99.99
57.98
6.905
0.003
22.49
1.183
0.007
19.53
1.063
Optimal technological improvement rate and infrastructure
99.99
57.37
6.865
0.003
23.46
1.247
0.007
19.17
1.040
No technological improvement and bad infrastructure
99.99
64.94
7.269
0.003
9.52
0.514
0.007
25.54
1.369
79
7.1.3. Recharging Infrastructure Based Scenarios (Scenario 3)
Recharging infrastructure is considered as an important criterion for customers
[15,27,28] because insufficient number of refueling point leads drivers to be stranded, or
stand in a queue for long hours for recharging. However, future station number and
adequacy of these stations are uncertain and these uncertain factors may affect the BEV
penetration. Hence, two distinct scenarios are constructed to evaluate impact of
infrastructure on BEV diffusion. First one is called as excellent infrastructure while second
one is called bad infrastructure. Excellent infrastructure represents a situation that the
number of recharging points would be sufficient to cover the charging demand. Thus, there
would be no time loss coming from waiting in a queue or being stranded. On the other
hand, in the case of bad infrastructure, number of recharging points would not be adequate
for BEV drivers and they would have to wait long hours for charging. Comparison of the
results of these two scenarios is given in Figure 7.19.
25
1
18.75
percent
1
1
12.5
2
1
2
1
6.25
1
1
0
1 2
2012
1 2
12
2016
12
12
2020
1
2
2
2
2
2
2024
2028
Time (year)
BEV fleet market share : excellent_infrastructure
BEV fleet market share : bad_infrastructure
2
2
2
1
1
1
2
2032
1
2
2036
1
2
1
2
2040
1
2
1
2
1
2
2
Figure 7.19. BEV market share under different infrastructure conditions.
The results show that recharging infrastructure is a relatively effective factor on BEV
penetration. For instance, according to the simulation results, the BEV fleet share would
reach 12.5% by 2034 in the excellent infrastructure scenario, whereas it reaches to same
penetration rate, 12.5%, by 2042 in the bad infrastructure scenario. Thus, insufficient
infrastructure causes the BEV diffusion to be delayed. Apart from that, cumulative CO2
80
reduction would reach around 17.8x106 tons in the excellent infrastructure scenario, while
it reaches around 14x106 tons in the bad infrastructure scenario by 2042.
The summarized results of recharging infrastructure related scenario are given in
Table 8.3.
Table 7.3. The results of recharging infrastructure related scenarios.
Market Share
2012 (%)
Base Run
CV
BEV
HEV
99.99
0.003
0.007
CV
BEV
HEV
99.99
0.003
0.007
CV
BEV
HEV
99.99
0.003
0.007
Market Share
2042 (%)
(Million)
59.47
19.76
20.77
Excellent infrastructure
58.94
20.56
20.50
Bad infrastructure
63.14
12.85
24.01
6.987
1.036
1.129
6,949
1.110
1.093
7.176
0.684
1.292
7.1.4. Introducing only BEV to the Market (Scenario 4)
As stated earlier, the model includes three types of vehicles; CV, HEV, and BEV.
However, there has been a debate about the possibility of HEV inhibiting BEV diffusion.
In addition to this debate, people also argue that How much CO2 reduction would be if
only BEVs were introduced to the market. Therefore, in this scenario, the model is
reconstructed considering imaginary world that only CVs and BEVs are available in the
market.
81
Fleet market shares of vehicles
2
100
4
4
2
4
2
4
2
4
2
4
2
4
2
percent
4
2
75
4
4
2
4
2
2
50
25
0
1 3
2012
1
3
3
1
2016
3
1
3
1
2020
3
1
3
1
2024
2028
Time (year)
CV fleet market share : base run 2
BEV fleet market share : Scenario 4
CV fleet market share : Scenario 4
1
2032
1
2
3
4
1
2
3
4
2036
1
2
3
4
2040
1
2
1
2
3
4
1
1
1
1
3
3
3
3
1
2
3
4
3
4
4
Figure 7.20. Fleet market share under Scenario 4.
Reduction of CO2
20
1
2
12
15
percent
1 2
12
10
1 2
12
1 2
5
1 2
0
12
2012
1 2
12
2016
1 2
12
2020
Reduction of CO2 : base run
Reduction of CO2 : Scenario 4
12
1 2
2024
2028
Time (year)
1
1
2
1
2
2032
1
2
1
2
2036
1
2
1
2
2040
1
2
1
2
1
2
Figure 7.21. Reductions of CO 2 under the base run vs Scenario 4.
As is seen from the graphs, if there were only C Vs and BEVs in the market, market
share of BEV would become higher than its market share in the base run. However, Figure
7.21 shows that if there were only CVs and BEVs in the market, reduction of CO 2 level
would be almost same compared to the base run. This is because of the fact that even fleet
market share of BEV increases; most of potential HEV customers prefer CVs over BEVs
and the more CV causes the more gas emissions.
82
The summarized results of this scenario are given in Table 7.4.
Table 7.4. The results of introducing only BEV to the market
Reduction of
Market Share
Market Share
Total sales until
CO
2012 (%)
2042 (%)
2050 (Million)
2 at 2042 (%)
Base Run
CV
BEV
HEV
Scenario 4
CV
BEV
99.99
59.47
6.987
0.003
19.76
1.036
0.007
20.77
1.129
Introducing only BEV to the market
99.999
71.67
7.670
0.001
28.33
1.482
17.32
17.39
7.1.5. Word of Mouth Related Scenarios (WoM) (Scenario 5)
A customer intending to buy a vehicle needs to be aware of vehicle types to take
them into her/his choice set. The awareness about new type of vehicles is provided via
marketing and word of mouth. Marketing influence will be analyzed in the policy section.
However, word of mouth related scenarios are analyzed in this section.
As mentioned before, the term, word of mouth (WoM), covers all actions of people
that help to spread information about EVs and actions that cause people to recognize EVs
in the study. For example, driving EV on the road, talking about EVs, mentioning about
them in the social media are included in the word of mouth definition in this study. WoM
is regarded as a substantial factor on innovation penetration. There are two WoM related
scenarios. In the first scenario (Scenario 5_1), influence of WoM of non-EV drivers, and in
the second one influence of non-EV drivers are intensified (Scenario 5_2).
7.1.5.1. Intensive non-EV Drivers‟ Word of Mouth (Scenario 5_1): In the Scenario 5_1,
word of mouth of non- EV drivers is analyzed. It should be noted that non-drivers represent
non-EV drivers who are aware of EVs in the study. In this scenario, exposure level coming
from non- EV drivers is doubled. In other words, it is assumed that they are twice as
effective as in the base run in terms of creating awareness about EVs. Behaviors obtained
from Scenario 5_1 is given below.
83
Sales market share
40
percent
4
4
30
4
4
20
10
4
4
0 2
1
2012
34 1
2 3
1
2016
3 4
4
2
1
2
3 4
1 2
1
2
1
1
2
1
2
2 3
1
2
2
3
3
3
3
4
3
3
1
2020
2024
2028
Time (year)
2032
2036
2040
1
1
1
1
1
1
1
2
2
2
2
2
2
3
3
3
3
3
3
4
4
4
4
4
4
Fleet market share
30
4
4
22.5
percent
4
4
15
41 2 34 1 23
2016
4
1
2020
2 3
1
1
1
2
1
1
1
2
2024
2028
Time (year)
1
2
2
2
2
3
4
0
1 23
2012
3
4
7.5
2
3
3
3
4
3
3
1
1
2
1
2
3
1
2
3
4
2040
1
2
3
4
2036
1
2
3
4
2032
3
4
1
2
3
4
4
Figure 7.23. Fleet market share of BEV and HEV under the Scenario 5_1.
As seen in the graphs, WoM of non-adopters has a remarkable impact on the
diffusion of BEVs and HEVs, particularly between 2016 and 2038. This range is mainly
because of the fact that in the first years of diffusion, number of non-EV users, who are
aware of EV is very low. Thus, even if their effectiveness is intensified, their impact
remains quite small. However, when their number increases, then information about EVs
begin to spread rapidly. After 2038, influence of WoM on the EV market share begins to
decline because the number of people who are not familiar with EV becomes considerably
84
lower. If unaware people diminish, then WoM naturally does not cause huge number of
people to gain awareness about EVs. In addition, cumulative CO 2 reduction would be
around 25 x 106 tons by 2042.
7.1.5.2. Intensive EV Drivers‟ Word of Mouth (Scenario 5_2): In the Scenario 5_2, word
of mouth of EV drivers is analyzed. In this scenario, exposure level coming from EV
drivers is doubled. In other words, they are twice as influential as in the base run in terms
of creating awareness. Behaviors obtained from Scenario 5_2 is given below.
Sales market share
35
4
percent
3
4
26.25
17.5
3
4
4
4 1
0 1 23 4 1 23
2012
2016
23
3 4
4
1
2
2
12
1
2
1
2
1
2 3
4
8.75
1
2
3
4
3
1
23
1
2020
2024
2028
Time (year)
2032
2036
2040
1
1
1
1
1
1
1
2
2
2
2
2
2
3
3
3
3
3
3
4
4
4
4
4
4
Fleet market share
25
4
percent
18.75
3
4
12.5
4
2
4
6.25
0
1 23 4 1 23 4
2012
2016
3
4
123
4 12 3
2020
4
1
1
1
1
2 3
2024
2028
Time (year)
2
1
2
1
2 3
4
3
3
2
1
2
1
1
2
1
2
3
1
2
3
4
2040
1
2
3
4
2036
1
2
3
4
2032
3
4
1
2
3
4
4
Figure 7.25. Fleet market share of BEV and HEV under the Scenario 5_2.
85
As can be seen from the graphs, WoM of the EV adopters is also influential on EV
diffusion. Its impact on market shares is similar to effect of non-adopters but it seems that
effect of adopters is less than effect of non-adopters. This is because of the fact that nonEV drivers outnumber EV drivers. In addition, cumulative CO 2 reduction would be around
19.7 x 106 tons by 2042.
The summarized results of WoM scenarios are given in Table 7.5.
Table 7.5. The results of WoM scenarios.
Market Share
2012 (%)
Base Run
CV
BEV
HEV
Scenario 5_1
CV
BEV
HEV
Scenario 5_2
CV
BEV
HEV
Market Share
2042 (%)
(Million)
99.99
59.47
6.987
0.003
19.76
1.036
0.007
20.77
1.129
Influence of non-EV drivers‟ word of mouth
99.99
49.73
6.245
0.003
24.27
1.371
0.007
26.00
1.536
Influence of EV drivers‟ word of mouth
99.99
54.70
6.696
0.003
22.00
1.171
0.007
23.30
1.285
7.1.6. Repurchasing Rate (Scenario 6)
As stated earlier, new cars are sold in the case of repurchasing or market growth.
Therefore, sales of any type of vehicles highly depend on how frequently people renew
their cars. However, long life of vehicles or desire to use same car for a long time or high
car prices may cause very low repurchasing. Regarding this situation, discard period of
vehicles is rearranged in the two scenarios to assess influence of repurchasing rate on EV
diffusion. Base run value of the average discard period is 8% per year which translates to
12.5 years of average life time. In the first scenario, discard period is set to 6% per year
(less repurchasing rate) that is equal to 17 years. In the second scenario, it is set to 10% per
86
year (high repurchasing rate) that is equal to 10 years. Influence of these two modifications
on the fleet shares are shown in Figure 7.26 and 7.27.
25
3
3
18.75
percent
3
3
12.5
3
31
6.25
0
1 23 1 2 31 2
2012
2016
31
31 2 3 12
2020
3 12
3 12
31 2
1
3
2
1
2
3
2
2
2032
1
2
2
2
1
2
2024
2028
Time (year)
1
BEV fleet market share : low_repurchase
BEV fleet market share : high_repurchase
1
1
1
2036
1
2
3
1
2
3
2040
1
2
3
1
2
3
1
2
3
3
Figure 7.26. BEV fleet market share under the re-purchasing scenario.
25
3
3
18.75
percent
3
3
12.5
3
31
6.25
0
1 23 1 2 31 2
2012
2016
1
3
312
3 12
2020
3 12
3 12
31
2
1
2
2
1
2
3
2
1
2032
1
2
2
2
1
2
2024
2028
Time (year)
1
HEV fleet market share : low_repurchase
HEV fleet market share : high_repurchase
1
1
1
2
3
2036
1
2
3
1
2
3
2040
1
2
3
1
2
3
3
Figure 7.27. HEV fleet market share under the re-purchasing scenario.
The results show that repurchasing rate is notably influential on the EV diffusion.
For example, BEV fleet share reaches 12.5% at 2033 in the case of high repurchasing.
However, it achieves to reach same penetration level (12.5%) at 2037 in the availability of
87
low repurchasing rate. If repurchasing rate is low in a city, then EV diffusion may be
delayed.
The summarized results of repurchasing scenarios are given in Table 7.6.
Table 7.6. The results of repurchasing scenarios.
Market Share
2012 (%)
Base Run
CV
BEV
HEV
Scenario 5_1
CV
BEV
HEV
Scenario 5_2
CV
BEV
HEV
99.99
0.003
0.007
99.99
0.003
0.007
99.99
0.003
0.007
Market Share
2042 (%)
59.47
19.76
20.77
Low repurchasing
64.98
17.03
17.99
High repurchasing
55.21
21.88
22.91
(Million)
6.987
1.036
1.129
5.961
0.788
0.858
8.003
1.288
1.406
7.2. Policy Analysis
When launching a new technology, various strategies are applied by governments
and manufacturers in order to provide faster diffusion of new technology. It is commonly
indicated that policies and incentives are needed to sustain broader penetration of new
technologies. For this reason, ten policies about EV diffusion are analyzed in this section.
Their main contexts are subsidy, tax, and marketing. The results of these policies will be
presented mostly by comparing them with the base run to provide a better insight. All
policies will be explained in detail in this section.
7.2.1. Subsidy Based Policies (Policy 1)
Most automobile manufacturers and researchers claim that financial incentive is
necessary for successful EV adoption. Subsidy for a purchase price is considered as one of
the financial incentive options [17,39,43]. Impacts of subsidy strategies are assessed with
88
the help of six subsidy policies. There are two basic different strategies are defined related
to subsidy regimes. Firstly, each one is app lied to only BEVs, and then for only HEVs.
Afterwards, they are applied to both BEVs and HEVs. It should be noted that there is no
clear information about subsidy regimes for EVs in Turkey. Therefore, subsidy strategies
are determined considering the work of Shepherd, Bonsall, and Harrison, which analyzes
EV diffusion in The UK [39]. Besides, there is no subsidy in the base run.
7.2.1.1. 5000 TL Subsidy for BEV (Policy 1_1_1): In this policy, it is assumed that every
BEV buyer would take 5000 TL as subsidy and this regulation would last until the end of
simulation. It is important to mention that this subsidy regulation is specific for only BEV
in the Policy 1_1_1. Impact of this strategy on the BEV fleet share is given in Figure 7.28.
25
2
percent
18.75
1
1
12.5
1
6.25
0
1 2
2012
1 2
12
2016
12
12
2020
1
12
1
2
1
2
1
2
2024
2028
Time (year)
BEV fleet market share : Policy 1_1_1
1
1
2
1
2
2
2036
1
2
2
2
2032
1
2
1
2
1
2
2040
1
2
1
2
1
2
Figure 7.28. Fleet market share of BEV under the Policy 1_1_1.
It can be obviously seen from the Figure 7.28 that „5000 TL subsidy for BEV‟
regime has little impact on the BEV fleet market share. In addition, total cost coming from
this subsidy is estimated as 5315 million TL.
7.2.1.2. 10000TL Subsidy for the First 10 Years for BEV (Policy 1_2_1): In this policy, it
is assumed that every BEV buyer would take 10000 TL for subsidy and the time duration
of this regulation would be the first ten years of the simulation. After the first ten years, the
89
subsidy would be removed because the government would less likely decide 10000 TL
subsidy for every vehicle throughout 30 years due to financial concerns. This subsidy
policy is applied to only BEV in the Policy 1_2_1. Impact of this strategy on BEV fleet
share is given in the Figure 7.29.
20
1
1
percent
15
1
1
10
1
1
1
5
0
12
2012
1 2
12
2016
1 2
12
2020
1
2
1
1
2
1
2
2
2
2
2
1
2
2
2
2024
2028
Time (year)
BEV fleet market share : Policy 1_2_1
2
2032
1
2
1
2
2036
1
2
1
2
2040
1
2
1
2
1
2
Figure 7.29. Fleet market share of BEV under the Policy 1_2_1.
As is seen from the Figure 7.29, „10000 TL subsidy for BEV regime‟ has ignorable
impact on BEV fleet market share. Moreover, total cost coming from this subsidy is
estimated as 518 million TL.
7.2.1.3. 5000 TL Subsidy for HEV (Policy 1_1_2): In the Policy 1_1_2, 5000 TL subsidy,
which is assessed in the Policy 1_1_1, is applied to only HEV instead of BEV. In other
words, it is assumed that every HEV buyer would take 5000 TL for subsidy and this
regulation would last until the end of simulation. This subsidy policy is applied to only
HEV in the Policy 1_1_2. Impact of this strategy on the HEV fleet share is given in Figure
7.30.
90
25
percent
18.75
1
12.5
1
1
6.25
0
1 2
2012
1 2
12
2016
12
12
2020
1
12
12
1
2
1
1
2
1
2
2
2
2032
1
2
2
2
2024
2028
Time (year)
HEV fleet market share : Policy 1_1_2
1
2
1
2
2036
1
2
1
2
2040
1
2
1
2
1
2
Figure 7.30. Fleet share of HEV under the Policy 1_1_2
As is seen in the Figure 7.30, „5000 TL subsidy for HEV regime‟ has little impact on
HEV sales. Moreover, total cost coming from this subsidy is estimated as 5120 million TL.
7.2.1.4. 10000TL Subsidy for 10 Years for HEV (Policy 1_2_2): In the Policy 1_2_2,
5000 TL subsidy, which is assessed in the Policy 1_1_2, is applied to only HEV instead of
the BEV. In this policy, it is assumed that every HEV buyer would take 10000 TL for
subsidy and the time duration of this regulation would be the first ten years of the
simulation. After the first ten years, the subsidy would be removed. This subsidy policy is
applied to only HEV in the Policy 1_2_1. Impact of this strategy on the HEV fleet share is
Figure 7.31 shows that „10000 TL subsidy for HEV regime‟ has quite a little impact,
almost zero, on HEV penetration. Moreover, total cost coming from this subsidy is
estimated as 788.56 million TL.
91
25
1
percent
18.75
1
1
12.5
1
1
6.25
0
1 2
2012
1 2
12
2016
12
12
2020
1
12
12
1
2
1
1
2
1
2
2
2
2
2
2024
2028
Time (year)
HEV fleet market share : Policy 1_2_2
2
2
2032
1
2
2036
1
2
1
2
2040
1
2
1
2
1
2
Figure 7.31. Fleet market share of HEV under the Policy 1_2_2.
7.2.1.5. 5000 TL Subsidy for Both BEV and HEV (Policy 1_3): In this policy, 5000 TL
subsidy is performed for both BEV and HEV to estimate if implementing 5000 TL subsidy
for both types is profitable or not. In this policy, it is assumed that every BEV and HEV
buyer would take 5000 TL for subsidy and this regulation would last until the end of
simulation. This subsidy policy is applied to both BEV and HEV in the Policy 1_3. Impact
of this strategy on the fleet shares is given in Figure 7.32.
As is seen in Figure 7.32, „5000 TL subsidy for both BEV and HEV regime‟ has
quite a little impact, almost zero, on both BEV and HEV penetration. Besides, total cost
coming from this subsidy is estimated as 10980 million TL. This is the highest cost among
all subsidy scenarios analyzed in the study.
92
Fleet market share
25
4
18.75
percent
4
4
12.5
4
6.25
0
1 23 4 1 23 4
2012
2016
123
4 12 3
2020
41
2 3
4
1
2 3
41
23
2024
2028
Time (year)
BEV fleet market share : Policy 1_3
HEV fleet market share : Policy 1_3
1
1
2
23
1
3
3
2036
2040
1
2
3
4
1
1
1
2
4
2
1
1
2032
2
23
23
1
2
3
4
1
2
3
4
3
4
4
Figure 7.32. Fleet market share of BEV and HEV under the Policy 1_3.
7.2.1.6. 10000TL Subsidy for 10 Years for Both EVs (Policy 1_4): In this policy, 10000
TL subsidy is performed for both BEV and HEV to estimate if implementing 10000 TL
subsidy for both types is profitable or not. Similar to the previous 10000 TL subsidy
regimes, it is assumed that every EV buyer would take 10000 TL and the duration o f this
regulation would be the first ten years of the simulation. After the first ten years, the
subsidy would be removed. This subsidy policy is applied to both BEV and HEV in the
Policy 1_4. Impact of this strategy on the fleet shares is given in the Figure 7.33.
Fleet market share
25
percent
18.75
2
4
12.5
2
4
6.25
23
2 41
1 23 4 1 23 4 1 3
0
2012
2016
2020
41
2 3
4
1
2 3
41
23
2024
2028
Time (year)
1
1
2
3
2036
1
1
2
3
4
2040
1
2
3
4
3
1
3
2032
2
3
4
23
2
1
1
2
4 1
41
3
4
1
2
3
4
4
93
Figure 7.33 shows that „1000 TL subsidy for both BEV and HEV regime‟ has little,
almost ignorable, impact on EV penetration. Moreover, total cost coming from this subsidy
is estimated as 1282 million TL.
The summarized results of subsidy-based policies are given in Table 7.7.
Table 7.7. The results of subsidy based policies.
Market
Share 2012
(%)
Base Run
CV
BEV
HEV
Policy 1_1_1
CV
BEV
HEV
Policy 1_2_1
CV
BEV
HEV
Policy 1_1_2
CV
BEV
HEV
Policy 1_2_2
CV
BEV
HEV
Policy 1_3
CV
BEV
HEV
Market Share
2042 (%)
Total Sales
until 2050
(Million)
99.99
59.47
6.987
0.003
19.76
1.036
0.007
20.77
1.129
5000 TL Subsidy for BEV for 30 years
99.994
59.2
6.972
0.001
20.26
1.063
0.005
20.54
1.117
10000 TL Subsidy for BEV for 10 years
99.994
59.43
6.984
0.001
19.80
1.040
0.005
20.77
1.128
5000 TL Subsidy for HEV for 30 years
99.994
59.16
6.970
0.001
19.54
1.024
0.005
21.30
1.158
10000 TL Subsidy for HEV for 10 years
99.994
59.43
6.982
0.001
19.76
1.035
0.005
20.81
1.134
5000 TL Subsidy for both BEV and HEV for 30
years
99.994
0.001
0.005
58.9
20.03
21.07
6.956
1.051
1.145
Total social
cost
(Million TL)
-
5315
518
5120
788.56
10980
94
Table 7.7. The results of subsidy based policies (cont.)
Policy 1_4
CV
BEV
HEV
10000 TL Subsidy for both BEV and HEV for
10 years
99.994
59.39
6.979
0.001
19.80
1.040
0.005
20.81
1.133
1282.12
7.2.2. Tax Based Policy (Policy 2)
A car is one of the product types that have private consumption tax (PCT) in Turkey.
However, amount of this tax varies depending on vehicle properties such as motor type, or
capacity. When representative vehicle types used in this work is considered, it is 37% for
internal combustion engine based vehicles. Besides, PCT of vehicles that have only electric
motor is 3% for incentive for EVs. On the other hand, hybrid vehicles, which are one of the
electric vehicles, have 37% private consumption tax since if a car have both internal
combustion engine and electric motor, then PCT for that car becomes 37% in accordance
with the law. However, there have still been debates in Turkey about that PCT should be
the 3% for also HEV to increase the EV penetration. In this respect, private consumption
tax of HEV is set to the 3% instead of 37% to analyze its effect on the EV diffusion and to
assess the necessity of tax regulations.
Figures show that new tax regulation may be supportive for HEV diffusion, but it
does not cause considerable change in the first years of diffusion due to low level of
awareness about EV. Figures show that application of new tax regulation after 2023 may
be profitable since tax regulation increase its influence due to upward awareness level.
95
HEV sales market share
40
30
2
2
1
percent
2
20
1
2
1
2
10
0 12
2012
1 2
12
1 2
2016
1
1
2
2020
1
2
1
2
2024
2028
Time (year)
HEV sales market share : Policy 2
1
2
1
1
1
1
2
2
2
1
2
2032
1
2
1
2
2036
1
2
1
2
2040
1
2
1
2
1
2
Figure 7.34. Sales market share of HEV under the Policy 2.
25
2
18.75
1
percent
2
1
2
12.5
1
2
1
2
1
2
6.25
0
1 2
2012
1
2
1 2
12
2016
12
12
2020
HEV fleet market share : Policy 2
1
12
1
2
2024
2028
Time (year)
1
2
1
2
1
2
2032
1
2
2036
1
2
1
2
2040
1
2
1
2
1
2
Figure 7.35. Fleet market share of HEV under the Policy 2.
96
The summarized results of the tax based policy are given in Table 7.8.
Table 7.8. The results of private consumption tax based policy.
Market
Share 2012
(%)
Base Run
CV
BEV
HEV
Policy 2
CV
BEV
HEV
99.99
0.003
0.007
99.994
0.001
0.005
Market Share
2042 (%)
59.47
19.76
20.77
Tax policy
58.61
19.12
22.27
Total sales until
2050 (Million)
6.987
1.036
1.129
6.938
1.002
1.211
7.2.3. Marketing Based Policies (Policy 3)
As mentioned before, people need social exposure to take EVs into their
consideration set. Two distinct social exposures are defined in the study. First one is word
of mouth of people and second one is marketing. Marketing is regarded as one of the most
efficient strategies to promote new products, or innovations. People recognize new
products and gain information about them thanks to marketing activities. Marketing
activities cover every advertisement channel such as TV, radio, newspaper, magazine, or
social media for this study. Impacts of marketing strategies on EV penetration are analyzed
via three policies. First policy helps to observe behavior if there were no marketing
activities about EVs. In the second policy, it is assumed that marketing activities are
lessened. In the final policy, it is assumed that marketing activities would continue for
limited duration and it would be stopped. In addition, it is important activities are also
included in the base run throughout simulation.
97
7.2.3.1. No Marketing Activities (Policy 3_1): In the Policy 3_1, it is assumed that there
would be no marketing activities about EVs. The effects of this case can be seen in Figure
7.36 and 7.37.
Sales market share
35
percent
26.25
1
2
1
2
8.75
1
2
1
2
2
0 1 234 1 34
2012
2016
1
3
1
2
3
34
34
3 4
3 4
2020
2024
2028
Time (year)
BEV sales market share : Policy 3_1
HEV sales market share : Policy 3_1
1
12
2
1
2
17.5
1
2
1
2
3
4
2032
1
2
3
4
2040
1
2
3
4
4
2036
1
2
3
3
4
1
2
3
4
4
1
2
3
4
3
4
4
Figure 7.36. Sales market share of BEV and HEV under the Policy 3_1.
Fleet market share
25
2
percent
18.75
2
1
2
12.5
1
2
1
2
6.25
1
2
1
2
0
1 23 4 1 23 4
2012
2016
1
123 41
1
2
2020
2024
2028
Time (year)
1
1
2
1
2
3
4
1
2
3
4
4
2040
1
2
3
4
2036
1
2
3
4
2032
1
2
3
34
34
34
34
3 4
3
34
3
4
4
As is seen from the graphs, if there would be no marketing activity for EV, adoption
process is heavily affected from this situation since people do not recognize EVs or they do
not know any information about them. Hence, they do not naturally take EVs into their
98
choice set, which results in low sales. The results also show that if marketing activities are
not implemented, the time to seize the fleet market share of 15% for total EVs is delayed
by more than a decade.
7.2.3.2. Less Marketing Activities (Policy 3_2):
In this policy, it is assumed that
marketing activities are halved. In other words, number of advertisements, interviews, or
news on both broadcast and printed media is cut in half. Impacts of this situation on the
market shares are given in the Figure 7.38 and Figure 7.39.
The simulation results show that if all marketing activities are halved, EV penetration
is affected negatively from this situation. However, it is still sufficient to make people
familiar with EVs.
Sales market share
35
percent
26.25
2
17.5
1
2
2
1
3
4
1
3
BEV sales market share : Policy 3_2
HEV sales market share : Policy 3_2
3
4
4
3
3
3
4
2020
12
2
3
4
1
2
3
4
4
4
1
2
8.75
0 1 23 4 1 23 4
2012
2016
1
2
1
1
2
2024
2028
Time (year)
1
1
2
1
2
3
1
2
3
4
2040
1
2
3
4
2036
1
2
3
4
2032
3
4
1
2
3
4
4
Figure 7.38. Sales market share of BEV and HEV under the Policy 3_2.
99
Fleet market share
25
2
percent
18.75
1
2
12.5
2
6.25
2
123 41
2
3
2020
41
3
4
1
3
3
4
2024
2028
Time (year)
1
1
2
3
1
2
3
4
2040
1
2
3
4
2036
1
2
3
4
2032
1
2
3
4
3
4
1
2
4
3
4
1
1
2
0
1 23 4 1 23 4
2012
2016
1
2
3
4
1
2
3
4
4
7.2.3.3. Marketing Activities for Limited Duration (Policy 3_3): In this scenario, it is
assumed that marketing activities continue for a while and then they are stopped. In this
context, influence of different marketing durations on the market shares is analyzed to
assess roughly optimum marketing duration. Hence, four different marketing periods are
determined that are 15, 10, 5, and 3 years. To illustrate, if the first period is regarded,
marketing activities will continue through 15 years from the beginning and then all
activities would be stopped at the end of 15th the year. The results of the limited marketing
duration are given in Figure 7.40 and Figure 7.41.
According to the simulation results, marketing is particularly important in the first
years of the diffusion process. If marketing activities are stopped before the 5th year of the
penetration process, certain amount of people would less likely recognize EVs or learn
information about them. As a matter of course, this situation causes low level of EV sales.
However, after 5 years, the number of people who are familiar with EVs sufficiently
increases to sustain adequate social exposure. Thus, marketing duration needs to exceed
minimum 5 years for EVs to be adequate for self- sustaining. Additionally, after 10 years,
marketing activities begin to lose its effect on the EVs sales since most people are already
aware of EVs.
100
According to the results, marketing for 3 years may result in around 12x106 tons
cumulative CO 2 reduction, while marketing for 15 years may result in around 17x10 6 tons
cumulative CO 2 reduction.
35
12
1 23 4 5
percent
26.25
12 34
1
8.75
2 3
4
5
3
12 45
0 23 4 5 12 3
1
2012
2016
BEV sales
BEV sales
BEV sales
BEV sales
BEV sales
5
1234
5
17.5
45
34 5
1 2
2020
2024
2028
Time (year)
2032
2036
2040
market share : base run
1
1
1
1
1
1
market share : marketing for 15 years 2
2
2
2
2
market share : marketing for 10 years
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
Figure 7.40. Sales market share of BEV under different marketing strategies.
20
1 23 4
percent
15
12
10
1
2 34
1
5
0
12
2012
3
12 4
12 34 5
34 51 23 45
2016
2020
23
4
3
4
5
5
5
5
3
1 2 45
2024
2028
Time (year)
2032
2036
2040
1
1
1
1
1
1
BEV fleet market share : marketing for 15 years
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
Figure 7.41. Fleet market share of BEV under different marketing strategies.
The summarized results of every marketing policy are given in Table 7.9.
101
Table 7.9. The results of different marketing policies.
Market Share
2012 (%)
Market Share
2042 (%)
Total sales until
2050 (Million)
99.99
59.47
6.987
BEV
HEV
Policy 3_1
0.003
0.007
19.76
20.77
No marketing activities
1.036
1.129
CV
BEV
HEV
99.994
0.001
0.005
84.85
7.383
7.768
8.475
0.327
0.350
Base Run
CV
Policy 3_2
CV
BEV
HEV
Policy 3_3
CV
BEV
HEV
Policy 3_3
CV
BEV
HEV
Policy 3_3
CV
BEV
HEV
Policy 3_3
CV
BEV
HEV
99.994
0.001
0.005
Less marketing activities
65.54
99.994
0.001
0.005
16.83
17.63
Marketing for 15 years
60.2
19.41
20.39
61.22
18.93
19.85
99.994
0.001
64.08
17.56
99.994
0.001
0.005
0.005
7.41
0.838
0.904
7.022
1.019
1.111
7.088
0.989
1.075
7.284
0.898
0.971
99.994
18.36
66.92
0.001
0.005
16.18
16.90
0.808
0.871
7.473
102
7.3. Combination of Scenario and Policies
Some scenarios and policies are examined in the previous two sections. In this
section,
three different scenario-policy combinations
and one scenario-scenario
combination are developed. Some scenarios or policies help to increase HEV penetration
while some positively affect BEV penetration. They are combined to analyze their
influences under situations different from the base run. In the first combination, marketing
influences are assessed under high electricity price scenario. In the second one, combined
effect of high gasoline price and bad infrastructure is analyzed. In the third combination, it
is assumed that the BEV technology is improved at optimal level but there are no
marketing activities for EVs. Lastly, tax regulation for HEV and advanced battery
improvements is analyzed together.
7.3.1. High Electricity Price and Over Marketing Activities (Combination 1)
In this combination, marketing influence is examined under the Scenario 1_3. As
stated in the Scenario 1_3, electricity cost increases and after a while, it becomes
considerably higher than gasoline cost. In this situation, BEV sales begin to reduce.
Besides, marketing policies in the Section 7.2.3 shows that marketing is a substantial
policy to increase BEV diffusion rate. In this scenario, patterns o f electricity and gasoline
costs are assumed to be identical patterns in the Scenario 1_3. In addition, the intensity of
marketing activities is doubled. In other words, number of advertisements, interviews on
the both broadcast and printed media are doubled. Effects of this combination are given in
Figure 7.42.
Figure 7.42 shows that if the electricity cost reaches greater values compared to the
gasoline cost, even marketing, which is one of the most effective strategies to accelerate
EV penetration, cannot stop fall of the BEV sales. Because marketing only provides people
to be familiar with EVs. However, if customer is not satisfied with EV attributes such as
operating cost, then they do not prefer EVs.
103
Sales market share
45
2
2
2
33.75
percent
2
2
22.5
2
2
11.25
2
2
0 12
2012
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
1
1
2016
2020
2024
2028 2032 2036
Time (year)
BEV sales market share : Combination 1
HEV sales market share : Combination 1
1
1
2
1
2
2040
1
2
2044
1
2
1
2
2048
1
2
1
2
Figure 7.42. Sales market share of BEV and HEV under the Combination 1.
7.3.2. High Gasoline Cost and Bad Recharging Infrastructure (Combination 2)
In this combination, it is assumed that electricity and gasoline costs behave exactly
same as the ones in Scenario 1_4. In other words, electricity cost increases normally (9%
annually), while gasoline cost rises rapidly. Moreover, it is assumed that there is bad
recharging infrastructure that is same as the bad infrastructure mentioned in the Scenario 3.
Under these conditions, sales market shares of BEV and HEV are analyzed. The results of
this combination and the Scenario 1_4 are given in the Figure 7.43.
The Scenario 1_4 results show that if gasoline cost begins substantially higher than
electricity cost, then after a while, the BEV sales exceed the HEV sales. However, the
Combination 2 results show that even the cost gap occurring between gasoline and
electricity becomes too much, the HEV sales become notably higher than the BEV sales
without sufficient recharging points.
104
Sales market share
40
4
percent
30
4
1 2
4
20
4
10
0 2
1
2012
4
2 4 1
34 1 3
2016
2
4
3
2
1
2020
3
3
3
3
3
3
1
2
1
2
2
2
2
1
2
4
1
4
41
1
3
3
2024
2028
Time (year)
BEV sales market share : Combination 2
2032
1
2
1
2
3
1
2
3
4
1
2
3
4
2040
1
2
3
4
2036
3
4
1
2
3
4
4
7.3.3. Advanced Improve ment and No Marketing (Combination 3)
In this section, the Scenario 2_2 and the Policy 3_1 are combined. In other words, it
is assumed that technological improvements would be advanced level and there would be
no waiting induced by the recharging infrastructure. However, no marketing activity is
applied in this combination. Under these conditions, the BEV sales market share and the
HEV sales market share are analyzed. The results of this combination are given in the
Figure 7.44.
As is seen from the Figure 7.44, even if technology is improved at advance level and
obstacles related to infrastructural are eliminated, penetration of BEV is delayed by about a
decade without marketing activities.
105
Sales market share
35
percent
26.25
1
2
1
2
1
4
3
3 4
2020
3 4
4
34
1
2
3
1
2
8.75
12
2
1
2
17.5
0 1 23 4 1 23 4
2012
2016
1
2
1
2
34
34
2024
2028
Time (year)
34
2032
2036
2040
1
1
1
1
1
1
1
2
2
2
2
2
2
BEV sales market share : Combination 3 3
3
3
3
3
3
4
4
4
4
4
4
7.3.4. Tax Regulation for HEV and Optimal Progress for BEV (Combination 4)
In this section, the Scenario 2_2 and the Policy 2 are combined. To clarify, private
consumption tax (PCT) for HEVs is assumed to be 3% instead of 37%. This value is equal
to the PCT of BEVs. Moreover, improvement level of BEV is assumed to be at advanced
level. In addition, it is assumed that there would be no waiting induced by insufficient
recharging point. Under these conditions, sales market share of BEV and HEV are
analyzed. The results of this combination are given in the Figure 7.45 and the Figure 7.46.
The figures show that even if the BEV technology is improved at advance level and
the number of recharging points is adequate to meet the charging demand; new tax
regulation for HEVs does not cause important change on the HEV sales.
106
40
percent
3
12
3
3
20
3
3
10
3
3
3
30
1 2
12
1 2
1 2
12
1 2
1 2
3 12
0
12
2012
3
31 2 12
31 2
2016
3 12
31 2
2020
2024
2028
Time (year)
2032
2036
2040
1
1
1
1
1
1
1
1
1
BEV sales market share : Policy 2 2
2
2
2
2
2
2
2
2
BEV sales market share : Combination 4 3
3
3
3
3
3
3
3
Figure 7.45. Sales market share of BEV under the Combination 4.
HEV sales market share
35
2
2
percent
26.25
2 31
31
2
3 1
2
3 1
3
231
231
17.5
23
8.75
2 31
0 23 1 2 31
1
2012
2016
1
2 3
2020
23
1
23
1
1
2024
2028
Time (year)
2032
2036
2040
1
1
1
1
1
1
1
1
1
HEV sales market share : Policy 2 2
2
2
2
2
2
2
2
2
HEV sales market share : Combination 4 3
3
3
3
3
3
3
3
Figure 7.46. Sales market share of HEV under the Combination 4.
Results: The simulation results show that sales volumes of both BEV and HEV are
always lower than CV sales throughout the simulation period in the base case. Moreover,
after three decades, each of BEV and HEV shares in the sales reaches only around 30%
while CV sales manage to capture 40% of the market at 2042. Furthermore, BEVs and
HEVs reach respectively 19.76% and 20.77% shares in the fleet in Istanbul by 2042. There
are two main reasons why EV sales are lower than CV‟s even after 3 decades. Firstly,
although the percentage of potential customers increases gradually, there are still people
who do not recognize EVs. These unaware customers buy CVs due to the perception of the
unavailability of other choices. The number of unaware people is substantially high in the
107
first years of the simulation. Therefore, EV sales are also relatively lower at the beginning
of the simulation. As to the second reason, although BEVs and HEVs may display
advantages compared to CVs, some of attributes still remain less efficient than CVs‟. For
example, BEVs and HEVs are profitable in terms of operating cost, and emissions.
However, the driving range of BEVs is lower, and the refueling time of BEVs is longer
than CV. Also, the maintenance cost of BEVs is markedly higher compared to the CVs
throughout the simulation period. In addition, the purchase price of HEV is also higher
than both BEV and CV. So, even after familiarity with EVs greatly increases in the public,
the market shares of BEVs and HEVs still fail to reach the market share held by CVs due
to the perception that conventional vehicles have more preferable properties. Apart from
these, the sales market share of HEVs is slightly higher than the BEVs market share in
throughout majority of the simulation period. This means that certain attributes of HEV,
which are maintenance cost utility and time utility, are seen as more preferable than BEVs‟
from the viewpoint of the customers. However, in the last years of simulation, the BEV
finally begins to be more preferable compared to the HEV due to improvements about
battery technology.
Total number of every vehicle type within Istanbul increases in the first 10 years of
simulation. However, the sales market share of CVs persistently declines throughout the
simulation. The reason why the total number of CVs increases while its sales market share
decreases is the result of growth in the automobile market. In addition to this, after the first
decade, while the number of sold BEVs and HEVs keeps increasing, the number of CVs
begins to decline.
The percentage of potential EV customers has an S-shaped behavioral pattern. It
begins with a 1% potential EV customer among all drivers in Istanbul, and converges to
100% near the end of the simulation. It grows slowly in the beginning of the diffusion due
to the low number of adopters compared to high number of non-adopters. Moreover, the
majority of the population among the non-adopters does not have adequate knowledge
about EVs in the beginning of the penetration process. Therefore, information about EVs
spreads very slowly during the first few years. After a while, the percentage of potential
EV customers grows faster because of the increase of non-adopters who become familiar
with EVs, and the adopters. This portion of aware people who drive EVs on the road, talk
108
about them, or mention them on the internet and in conversation, this then leads to a
positive rise of further potential EV customers.
Another important point about EVs is the reduction level of greenhouse gas
emissions, which is one of the major reasons of why EVs are proposed as a necessary
replacement for CVs. The results indicate that the fleet market shares of BEVs and HEVs
reach 19.76% and 20.77%, respectively by 2042 in the base run. When this observation is
regarded, CO 2 reduction in the transportation sector reaches around 17.32% in 2042.
Moreover, cumulative CO 2 reduction reaches 17.07x106 tons by 2042.
The results also imply that both gasoline costs and electricity costs pose an influence
on vehicle sales, and thus EV diffusion. However, it must be noted that these influences are
mainly related to the driving cost gap between the use of gasoline and electricity. In other
words, even if electricity cost increases; this situation would not significantly affect BEV
sales; unless electricity cost comes close to gasoline cost or exceeds it. Similarly, the rapid
increase in gasoline cost would not heavily influence BEV and HEV sales unless the gap
between electricity and gasoline costs becomes notably large.
Moderate or optimistic improvements in battery technology can cause BEV sales
volumes to increase. In addition to this, moderate or optimistic technological
improvements ultimately lead to greater BEV sales over HEVs. However, despite of the
advanced improvements, BEV would achieve to reach 23.46% of the market by. In other
words, these progresses do not cause a significant raise of the BEV market share unless
battery technology keeps pace with or exceeds CV technology in terms of driving range,
refueling time, and maintenance cost. The results also show that even if there would be no
technological improvements, no fall in purchase price, and no progress in recharging time
for BEVs, the battery electric vehicles still succeed to capture 10% of the market. This
means that BEVs may survive with their current technology. Finally, if no progress is
made in BEV technology, this situation results in HEV sales to increase. We can deduce
from these consequences that although both BEVs and HEVs are categorized as electric
vehicles and CVs is supposed to be their competitor, these two distinct electrical vehicles
also compete with each other.
109
If the number of recharging points becomes sufficient to meet consumer electricity
demands, EV diffusion speeds up and captures a higher share in the market. If recharging
points are not adequate, BEV penetration is delayed. Therefore, the government should
prioritize to the construction of new recharging stations in order to increase BEV sales. It is
important to point out that inadequate charging infrastructure for BEVs results in HEVs
having a higher market share, this is not the case with sufficient infrastructure in place. In
other words, sufficient BEV recharging points cause HEV sales to decline. This situation
encourages the inference that BEVs and HEVs may inhibit each other sales.
The results show that if only BEVs are launched to the market, then the BEV fleet
share would likely be about 7.5% more than its base run value in 2042. However, in this
case, the CV fleet share becomes around 72%. This share is 12% more than the CV share
in the base run. Moreover, CO 2 reduction is about 0.07% higher than its value in the base
run. These results show that even though not launching HEVs to the market may be
profitable for BEV sales; this situation causes a higher rise in CV sales compared to a rise
in BEV sales. This is mainly due to the fact that most of potential HEV customers choose
CVs instead of BEVs. In addition, the amount of CO 2 released from CVs is greater than the
gas stemming from BEVs or HEVs. In this regard, it can be deduced that not introducing
HEVs to the market does not cause significant change to the climate, in terms of CO 2
reduction.
According to the simulation results, word of mouth (WoMs) of both EV drivers and
non-EV drivers has a remarkable impact on EV penetration. WoM influence strengthens
particularly between 2016 and 2038 because the number of EV users and aware non- EV
drivers are very low in the first years of diffusion. Thus, even if all of them talk about EVs,
drive them on the road, or mention them in the social media, their total impact st ill remains
quite small in the opening years. However, when the number of aware people increases,
the amount of exposure also increases. Hence, more people recognize EVs between 2016
and 2038. However, after 2038, the influence of WoM on EV market share be gins to
decline because the number of people who are not familiar with EV becomes considerably
lower. As unaware people diminish, WoM does not then cause a huge number of people to
gain awareness about EVs. Moreover, exposure coming from non-adopters may be more
110
influential compared to adopters of the technology due to a greater number of non- EV
drivers.
If the repurchasing rate increases in Istanbul, this situation would likely cause BEV
and HEV penetration to gain speed. The rise of repurchasing rates also affects CVs
negatively. This observation shows that policies aimed to increase the repurchasing rate
may accelerate both BEV and HEV diffusion.
Subsidies have a relatively small impact on the sales of both BEVs and HEVs in
Turkey. Because even with 5000 TL or 10000 TL subsidy, EV prices become higher
compared to the CVs throughout the majority of the simulation. In addition, there is low
number of potential EV customers in the beginning of the simulation. However, influence
of 5000 TL subsidy regime on EV sales begins to increase gradually but slightly after
2035. Because potential EV customers increase and EV prices come close to CV prices
due to both subsidies and decrease in EV price coming from learning by doing. Even so,
implementation of subsidy strategies may not be adequate alone to provide for a rise of the
market share of EVs. For this reason, subsidies do not show a considerable change in CO 2
reduction. In addition to its small impact, it is important to point out that subsidy regimes,
which are mentioned in the Policy section, result in a huge total cost. Particularly „5000 TL
for all EV subsidy-regime‟ may prove to be very costly.
If the private consumption tax (PCT) for HEVs is set to 3% instead of 37%, HEV
diffusion is positively affected by this tax regulation. HEV shares in the fleet in Istanbul
reach to 22.27% by 2042 while it reaches 20.77% in the base run. Therefore, even new tax
regulation induces HEV sales to increase, it does not cause substantial change. According
to the results, this tax regulation may be more effective if it is applied after 2023. This is
mainly due to the fact that the influence of this policy gradually increases in parallel with
the upward awareness level. Apart from that, the results indicate that when HEV sales go
up due to this policy, BEV sales become lower compared to the base run even if just by a
drop.
The results suggest that sufficient marketing activities are necessary to provide
successful and rapid EV penetration. Inadequate marketing levels cause the penetration of
111
EVs to be delayed more than one decade. According to the simulation results, marketing is
particularly important in the first 5 years of the diffusion process. If the marketing
activities were stopped before the 5th year, certain amounts of people would less likely
recognize EVs or learn information about them. This situation causes a low level of EV
sales. However, after 5 years, the number of people who are familiar with EVs sufficiently
increases to sustain an adequate level of social exposure. Thus, marketing duration needs
to exceed a minimum of 5 years for EVs to be adequately self-sustaining. Additionally,
after 10 years, marketing activities begin to lose their effect on EV sales since most people
are already aware of EVs. Thus, after 10 years, marketing activities can be stopped or their
level can be reduced to cut cost. In this regard, manufacturers should give importance to
their marketing activities and these activities should not be removed before the market
share is high enough to sustain a steady social exposure rate.
If WoM and marketing scenarios are considered together, it can be deduced that
marketing is particularly important in the first 5 years of penetration. In the same years,
WoM has a weak influence due to the small number of aware people. However, after 5
years, WoM get strong enough to sustain awareness without marketing.
If electricity costs reach greater values and the discrepancy between electricity and
gasoline costs becomes considerably high, even marketing, which is one of the most
effective strategies to accelerate EV penetration, cannot stop the fall of BEV sales. This is
because of the fact that marketing activities can only provide people to be fami liar with
EVs. However, if customers do not feel satisfied with EV attributes such as operating
costs, then they also do not prefer EVs. Similarly, even if gasoline costs grows increasingly
and electricity costs increases normally, then the number of people who prefer HEVs or
CVs instead of BEVs becomes higher when recharging points are insufficient.
Even if technological developments are progressed at an advanced level and
recharging points become sufficient to meet the demand, the penetration of BEVs is still
delayed by about a decade without marketing activities. The reason for this is that drivers
would less likely notice BEVs, and their advanced attributes without marketing activities;
no matter how advanced they are. If technological improvements of BEVs reach advanced
level and a recharging infrastructure becomes adequate for drivers, the 3% private
112
consumption tax (PCT) instead of the 37% PCT policy for HEVs has an ignorable impact
on HEV sales.
113
8. CONCLUSION
Electric vehicles (EVs) have been proposed to replace conventional vehicles due to
their potential advantages related to both the environment and energy consumption.
However, certain technical and social obstacles will come with the adoption of EV
technology. Besides, the eventual reduction of CO 2 emissions will heavily depend on the
diffusion trajectory of electric vehicles. With regards to this, the study investigates the
following two questions: Firstly, what are the plausible diffusion patterns of electric
vehicles for Istanbul under different scenarios developed considering both local and global
socio-economic, governmental, technological factors and their interaction with each other?
Secondly, what is the expected extent of the diffusion rate for Istanbul after three decades?
In the study, a dynamic simulation model is constructed by employing system
dynamics methodology. Subsequent to model construction, the model is validated with
structural and behavior tests. After the validation, various scenario and policy analysis are
performed.
It is observed that the sales volumes for EVs are always lower than the CV sales
throughout the simulation period, which is from 2012 to 2042, within the base case. After
three decades, both of the BEV and HEV shares in annual sales reaches only around 30%
while CV sales captures 40% of the market at 2042. There are two main reasons why the
EV sales market shares are lower than CVs‟ even after 3 decades. Firstly, although the
percentage of potential EV customers goes up gradually, there are still people in the market
who do not consider EVs as an option. These unaware customers directly buy CV due to
the perception of the unavailability of other choices. Secondly, although EVs may have
preferable sides compared to CVs, they still have some attributes that remain less efficient
than CVs‟. For instance, BEVs and HEVs are advantages in terms of operating cost, and
emission utility. However, the driving range of BEVs is lower, the refueling time of BEVs
is longer and also the maintenance cost of BEV is higher compared to CVs. These
attributes improves gradually but they fail to improve enough to capture higher market
share than CVs. In addition to these setbacks, the purchase price of HEVs is much higher
than both BEVs and CVs. Hence, even after familiarity with EVs significantly increases,
114
market share of neither BEVs nor HEVs succeeds to reach the market share of CVs, simply
due to the more preferable properties of conventional vehicles. Moreover, according to the
results, CO 2 reduction in the transportation sector would still only reach around 17% in
2042 and cumulative CO 2 reduction in Istanbul will be around 17.10 6 tons by 2042.
Gasoline costs and electricity costs have influence on EV diffusion. However, it is
important to point out that their impact on diffusion is mainly associated with a mobility
cost discrepancy between gasoline and electricity. For instance, even marketing, which is
regarded as one of the most influential strategies used to raise BEV sales, cannot stop BEV
sales to decrease if electricity cost exceeds gasoline cost and it continues to rise.
Furthermore, moderate or optimistic improvement of battery technologies would lead to
BEV shares in annual sales to increase. However, if battery technology does not keep pace
with or exceed CV technology in terms of driving range, refueling time, and maintenance
cost; then technological improvements cannot create a significant raise in the BEV market
share. Contrary to expectations, even if no technological improvements were realized,
BEVs would still succeed to penetrate around 10% of the market based solely on its
current technology within the 30- year span of the model. Apart from this, a sufficient
number of recharging points may lead to faster diffusion of BEV‟s as well, causing higher
fleet market share overall. For example, even if electricity cost increases normally and
gasoline cost grows increasingly as well, the number of consumers who prefer HEVs or
CVs over BEVs still becomes higher when recharging points are insufficient. Therefore,
the government should attach importance to new recharging point constructions if it wants
to ensure a successful BEV diffusion. If only BEVs are introduced to the market as an
electrical vehicle, both BEV and CV sales increase compared to the cases when all three
types are active in the market. However, this strategy causes a higher rise in CV sales over
BEV sales. For this reason, not introducing HEV to the market would be less likely to
create a significant drop in CO 2 emission compared to the market conditions when all three
types are available.
Two basic mechanisms help customers to gain awareness about EVs, which result in
being potential EV customer. These two mechanisms are marketing and word-of mouth
(WoM). Both marketing activities and word of mouth have a remarkable impact on rapid
EV diffusion. For instance, an inadequate marketing level causes significant penetration of
115
EVs into the market to be delayed by more than one decade even there would be optimistic
technological and infrastructural improvements about BEVs. In the first years of diffusion,
WoM has a weak influence due to the small number of aware consumers active in the
marketplace. Therefore, marketing activities should continue to spread information about
EVs in order to guarantee consumer recognition, particularly in the first years of the
diffusion process. However, after the first 5 years, the number of people who are familiar
with EVs increases and WoM becomes strong enough to sustain an adequate social
exposure without marketing. Moreover, after the 10th year of the penetration process, the
effectiveness of marketing activities on EV sales begins to decline since most of people are
already aware of EVs. So, after the 10th year, stopping marketing activities or reducing its
level would be profitable in terms of cutting the costs for the government and automobile
companies. Accordingly, the government and automobile companies should allocate an
adequate amount of their budgets to the marketing activities until the market share of EVs
is sufficient enough to be self-sustaining.
Policies that focus on raising the repurchasing rate may also result in faster EV
penetration. For instance, the government currently withdraws vehicles older than 20 years
from the market. This age can be decreased. Although subsidy regimes are proposed as an
effective way to speed up EV penetration by some manufacturers and academic authorities,
subsidies will have a small impact on the sales of both BEVs and HEVs in Turkey.
Because of this, subsidies would less likely create considerable change on CO2 reduction.
Apart from its low impact on sales, subsidy regimes also are likely to bring about a huge
overall cost. When the little impacts of subsidy regimes on diffusion process and CO 2
reduction are regarded, allocating this huge amount of money to subsidy regimes seems
unprofitable from the perspective of the society as a whole. Thus, profit and loss accounts
should be analyzed in detail by the government before implementation of any subsidy
regimes. Beside subsidies, setting the private consumption tax to 3% instead of 37% may
be supportive for HEV diffusion. However, it does not cause significant increase in HEV
sales, particularly in the first years of diffusion. Besides, if battery technology and a
recharging infrastructure for BEVs are improved at optimistic level, this tax regulation has
ignorable impact on HEV sales.
116
Finally, both BEVs and HEVs are categorized as electric vehicles; while CVs are
considered to be their basic competitor within the transportation sector. It is believed that
HEVs are crucial to attract CV customers in EV diffusion process. However, if HEVs are
supported too much by the government or manufacturers, this situation may inhibit BEV
penetration in Istanbul because BEVs and HEVs also compete with each other. Therefore,
after HEVs succeed to attract attention of CV customers, the government and automobile
firms may reduce or stop incentives for HEVs to provide broader BEV penetration.
As future research, different type of alternative fuel vehicles such as plug in hybrid
vehicles, and fuel cell vehicles, may be included in the model. Moreover, adding new
vehicle attributes may enrich the model. In the study, it is assumed that CVs do not have
technological progress throughout 3 decades. Improvement in CV technology can be
regarded in future researches. Finally, marketing is an exogenous variable in the study. It
may be turn into an endogenous variable with a good extension of the model.
117
APPENDIX A: MODEL EQUATIONS
AFV number per station=((total number of BEV)/(Number of
Recharging Stations of
BEVs))/Reference vehicle number per station
Units: Dmnl
BEV electricity unit cost= Electricity unit price
Units: TL/km
BEV discard rate=total number of BEV*BEV discard fraction
Units: vehicle/year
BEV
driving
range=190*("Smoothed
learning
ratio
for
R&D"Âlpha for BEV driving range)
Units: km
BEV fleet market share=
total number of BEV/(total number of CV+ total number of
BEV+total number of HEV)
Units: Dmnl
BEV emission utility for A= BEV emission rate*BEV weight of
emission utility for A
Units: Dmnl
BEV emission utility for B=BEV emission rate*BEV weight of
emission utility for B
Units: Dmnl
BEV operating cost utility for A=BEV electricity unit
cost*BEV weight of refueling cost for A+BEV maintenance
cost*BEV weight of maintenance cost for A
Units: Dmnl
BEV operating cost utility for B=BEV electricity unit cost*EV
weight of refueling cost for B+
BEV
maintenance cost*EV
weight of maintenance cost B
Units: Dmnl
BEV perceived utility for A=BEV emission utility for A+BEV
operating cost utility for A
+BEV purchase price utility for A+BEV time utility for A
Units: Dmnl
BEV perceived utility for B=BEV emission utility for B+BEV
operating cost utility for B
+BEV purchase price utility for B+BEV time utility for B
Units: Dmnl
BEV purchase price=BEV purchase price before taxes *(1+BEV
private consumption tax)*(1.18)
Units: TL
BEV purchase price before taxes=38261*(EV learning ratio for
costÂlpha forBEV purchase price)
Units: TL
BEV purchase price utility for A= (BEV purchase price/5000)/
118
(LN(BEV annual income level of household for A))*BEV weight
of purchase price A
BEV purchase price utility for B=(BEV purchase price/5000)/
(LN(Allocated money per car for B)) *BEV weight of purchase
price B
Units: Dmnl
BEV refueling time=(0.95)*("Smoothed learning ratio for
R&D"Âlpha for BEV refueling time)
Units: hour
BEV sales=BEV sales market share*market growth+"Total re purchase"*BEV sales market share
Units: vehicle/year
BEV sales market share= BEV sales market share for
A*percentage
of
group
A+BEV
sales
market
share
for
B*percentage of group B
Units: Dmnl
BEV sales market share for A=(EXP(BEV perceived utility for
A)/(EXP(CV perceived utility for A) +EXP(BEV perceived
utility for A)+EXP(HEV perceived utility for A)))*Percentage
of potential customers for EV
Units: Dmnl
BEV sales market share for B=(EXP(BEV perceived utility for
B)/(EXP(CV perceived utility for B) +EXP(BEV perceived
utility for B)+EXP(HEV perceived utility for B)))*Percentage
of potential customers for EV
Units: Dmnl
BEV sales market share in potential EV customers=percentage
of group A*BEV sms in potential EV customers A+percentage of
group B*BEV sms in potential EV customers B
Units: Dmnl
BEV sms in potential EV customers A= BEV sales market share
for A/Percentage of potential customers for EV
Units: Dmnl
BEV sms in potential EV customers B= BEV sales market share
for B/Percentage of potential customers for EV
Units: Dmnl
BEV time utility for A= (Max range/BEV driving range) *BEV
refueling time*Effect of infrastructure on BEV refueling time
Units: Dmnl
BEV time utility for B=(Max range/BEV driving range)*BEV
refueling time *Effect of infrastructure on BEV refueling
time*BEV weight of time B
Units: Dmnl
Current number of stations= Number of BEV stations planned to
be constructed + Number of Recharging Stations of BEVs
Units: station
construction=Municipality
criteria*Effect
of
desired
constraction on nmb of recharging stations
Units: station/year
119
CV fleet market share=total number of CV/(total number of CV
+total number of BEV+total number of HEV)
Units: Dmnl
CV discard rate=total number of CV*CV discard fraction
Units: vehicle/year
CV emission rate=1*("CV learning ratio for R&D"Âlpha for CV
emmision rate)
Units: Dmnl
CV emission utility for A=CV emission rate*CV weight of
Units: Dmnl
CV emission utility for B=CV emission rate*CV weight of
Units: Dmnl
CV number per refueling point=((total number of CV+total
number of HEV)/(Number of refueling stations of CV*CV number
of refueling point in every station ))/Reference CVnumber per
refueling point
Units: Dmnl
CV operating cost utility for A=CV gasoline unit cost*CV
weight of refueling cost for A
+CV maintenance cost*CV weight of maintenance cost for A
Units: Dmnl
CV operating cost utility for B=CV gasoline unit cost*CV
weight of refueling cost for B+CV maintenance cost*CV weight
of maintenance cost for B
Units: Dmnl
CV perceived utility for A=CV emission utility for A+CV
operating cost utility for A+CVpurchase price utility for
A+CV time utility for A
Units: Dmnl
CV perceived utility for B=CV emission utility for B+CV
operating cost utility for B+
CV purchase price utility for B+CV time utility for B
Units: Dmnl
CV purchase price=CV purchase price before taxes*(1+CV
Units: TL
CV purchase price before taxes=25000*(CV learning ratio for
costÂlpha for CV purchase price)
Units: TL
CV purchase price utility for A=(CV purchase price/5000)/
(LN(CV annual income level of household for A)) *CV weight of
purchase price for A
Units: Dmnl
CV purchase price utility for B=(CV purchase price/5000)/
(LN(CV Allocated money per car for B))*CV weight of pu rchase
price for B
Units: Dmnl
120
CV sales=CV sales market share *market growth+"Total re purchase"*CV sales market share
Units: vehicle/year
CV sales market share=Percentage of potential customers
CV*(CV SMS comes from PAFVC for
A*percentage of group A+CV
SMS comes from PAFVC for B*percentage of group
B+"CV
SMS
for A comes from non-PAFVC"+"CV SMS for B comes from non
PAFVC")
Units: Dmnl
CV SMS comes from PAFVC for A=
(EXP(CV perceived utility for A)/(EXP(CV perceived utility
for A) +EXP(BEV perceived utility for A)+EXP(HEV perceived
utility for A)))*Percentage of potential customers for EV
Units: Dmnl
CV SMS comes from PAFVC for B=
(EXP(CV perceived utility for B)/(EXP(CV perceived utility
for B) +EXP(BEV perceived utility for B)+EXP(HEV perceived
utility for B)))*Percentage of potential customers for EV
Units: Dmnl
"CV SMS for A comes from non-PAFVC"=
"Percentage of potential non-EV customers"*percentage of
group A
Units: Dmnl
"CV SMS for B comes from non-PAFVC"=
"Percentage of potential non-EV customers"*percentage of
group B
Units: Dmnl
CV time utility for A=(Max range/CV driving range)*CV
refueling time*
Effect of infrastructure on CV refueling time*CV weight of
time for A
Units: Dmnl
CV time utility for B=(Max range/CV driving range)*CV
refueling time*Effect of infrastructure on
CV
refueling
time*CV weight of time for B
Units: Dmnl
Desired constraction=Number of BEV stations planned to be
constructed/Construction delay
Units: station/year
Effect of desired constraction on nmb of recharging
stations=LOOKEXTRAPOLATE("Graph of desired
cons. on nmb of
rec. sta.", Desired constraction/Municipality criteria)
Units: Dmnl
Effect of infrastructure on BEV refueling time=LOOKUP
EXTRAPOLATE(Graph of effectof infrastructure on BEV refueling
time,AFV number per station )
Units: Dmnl
Effect of
infrastructure
on CV
refueling
time=LOOKUP
EXTRAPOLATE(Graph of
121
effect of queue on CV refueling time, CV number per refueling
point)
Units: Dmnl
Effect of infrastructure on HEV refueling time=
LOOKUP EXTRAPOLATE(Graph of effect of queue on HEV refueling
time, HEV number per refueling point)
Units: Dmnl
Electricity unit price=Current electricity unit price
Units: TL/km
Emission level of one HEV=Emission level of one CV*(1-HEV
improvement)
Units: g/(km*vehicle)
Emission level of total BEVs=Annual range of BEV*Emission
level of one BEV
*total number of BEV
Units: g/year
Emission level of total CVs=Annual range of CV*Emission level
of one CV*total number of CV
Units: g/year
Emission level of total HEVs=Annual range of HEV*Emission
level of one HEV
*total number of HEV
Units: g/year
EV
Customer
Awareness
Gain=Total
social
exposure*(1Percentage of potential customers for EV)
Units: Dmnl/year
EV customers awareness loss=Percentage of potential customers
for EV*EV awareness loss fraction
Units: Dmnl/year
EV learning ratio for cost="Total BEV sales (Total experience
in Istanbul)"/EV reference experience
Units: Dmnl
EV sale accumulation=BEV sales
Units: vehicle/year
EV total sales market share=BEV sales market share+HEV sales
market share
Units: Dmnl
expectation formation of EV=(total number of BEV-Expected
number of BEV in Istanbul)/
EV estimation time
Units: vehicle/year
expectation formation of HEV=
(total number of HEV-Expected number of HEV in Istanbul)/HEV
estimation time
Units: vehicle/year
Expected number of BEV in Istanbul= INTEG (expectation
formation of EV,1)
Units: vehicle
Expected number of HEV in Istanbul= INTEG (expectation
formation of HEV,1)
122
Units: vehicle
Gap=MAX(0, Desire number of BEV stations-Current number of
stations )
Units: station
"Graph of desired cons. on nmb of rec. sta."([(0,0)(1.6,1.5)],(-0.2,0),(0,0),(0.197248,0.243421),
(0.408257,
0.473684),(0.683486,0.736842),
(0.963914,0.907895),
(1.2156,0.980263),(1.5,1),(1.6,1))
Units: Dmnl
Graph of effect of infrastructure on BEV refueling time
([(1,0)(12.5,4)],(0,1),(1,1),(2.61774,1.38596),(4.05963,1.929
82),(5.04434,2.49123),(6.72783,3.12281),(8.40979,3.57895),
(9.78593,3.80702),(11.1009,3.9),
(12.5,4),(12.7523,4))
Units: Dmnl
Graph
of
effect
of
queue
on
HEV
refueling
time(
[(0,0)(5,1.5)],(0,1),(1,1),(1.66667,1.04386),(2.23242,1.09649
),(2.70642,1.20175),(3.18043,1.2
561),(4,1.3),(5,1.3))
Units: Dmnl
Graph
of
effect
of
queue
on
CV
refueling
time([(0,0)(5,2)],(0,1),(1,1),(1.66667,1.04386),(2.23242,1.09
649),(2.70642,1.20175),(3.21101,1.31579),(4,1.4),(5,1.4))
Units: Dmnl
HEV discard rate=total number of HEV*HEV discard fraction
Units: vehicle/year
HEV driving range=CV driving range*(1+HEV improvement)
Units: km
HEV emission rate=1*(1-HEV improvement)
Units: Dmnl
HEV emission utility for A=(HEV emission rate)*HEV weight of
Units: Dmnl
HEV learning ratio="Total HEV sales in Istanbul (total
experience)"/HEV initial experience
Units: Dmnl
HEV emission utility for B=(HEV emission rate)*HEV weight of
Units: Dmnl
HEV fleet market share=total number of HEV/(total number of
CV
+total number of BEV+total number of HEV)
Units: Dmnl
HEV gasoline prices=CV gasoline unit cost*(1-HEV improvement)
Units: TL/km
HEV number per refueling point=CV number per refueling point
Units: Dmnl
HEV operating cost utility for A=(HEV gasoline prices)
*HEV weight of refueling cost for A+HEV maintenance cost*HEV
weight of maintenance cost for A
Units: Dmnl
123
HEV operating cost utility for B=(HEV gasoline prices)
*HEV weight of refueling cost for B+HEV maintenance cost *HEV
weight of maintenance cost for B
Units: Dmnl
HEV perceived utility for A=HEV emission utility for A+HEV
operating cost utility for A
+HEV purchase price utility for A+HEV time utility for A
Units: Dmnl
HEV perceived utility for B=HEV emission utility for B+HEV
operating cost utility for B+HEV
purchase price utility for
B+HEV time utility for B
Units: Dmnl
HEV purchase price=HEV purchase price before taxes*(1+HEV
Units: TL
HEV purchase price=HEV purchase price before taxes*(1+HEV
Units: TL
HEV purchase price before taxes=36500*(HEV learning ratio^HEV
alpha)
Units: TL
HEV purchase price utility for A=(HEV purchase price/5000)/
(LN(HEV annual income level of household for A)) *HEV weight
of purchase price for A
Units: Dmnl
HEV purchase price utility for B=(HEV purchase price/5000)/
(LN(HEV Allocated money per car for B))*HEV weight of
purchase price for B
Units: Dmnl
HEV refueling time=CV refueling time
Units: hour
HEV sales=HEV sales market share*market growth+"Total re purchase"*HEV sales market share
HEV sale accumulation=HEV sales
Units: vehicle/year
HEV sales market share= HEV sales market share for
A*percentage of group A
+HEV sales market share for B*percentage of group B
Units: Dmnl
HEV sales market share for A=
(EXP(HEV perceived utility for A)/(EXP(CV perceived utility
for A) +EXP(BEV perceived utility for A)+EXP(HEV perceived
utility for A)))*Percentage of potential customers for EV
Units: Dmnl
HEV sales market share for B=
(EXP(HEV perceived utility for B)/(EXP(CV perceived utility
for B) +EXP(BEV perceived utility for B)+EXP(HEV perceived
utility for B)))*Percentage of potential customers for EV
Units: Dmnl
124
HEV sales market share in potential EV customers=percentage
of group A*HEV sms in potential
EV customers A+percentage
of group B*HEV sms in potential EV customers B
Units: Dmnl
HEV sms in potential EV customers A=
HEV sales market share for A/Percentage of potential
customers for EV
Units: Dmnl
HEV sms in potential EV customers B=
HEV sales market share for B/Percentage of potential
customers for EV
Units: Dmnl
HEV time utility for A=(Max range/HEV driving range)*HEV
refueling time
*Effect of infrastructure on HEV refueling time*HEV weight of
time for A
Units: Dmnl
HEV time utility for B= (Max range/HEV driving range)*HEV
refueling time
*Effect of infrastructure on HEV refueling time*HEV weight of
time for B
Units: Dmnl
Number of BEV stations planned to be constructed= INTEG
(planned construction
-construction ,0)
Units: station
Number
of
Recharging
Stations
of
BEVs=
INTEG
(construction,13)
Percentage of EV drivers= (total number of BEV+total number
of HEV)/total number of vehicle
Units: Dmnl
Percentage of potential customers for EV= INTEG (EV Customer
Awareness Gain-EV customers awareness loss,0.005)
Units: Dmnl
Percentage of potential EV customers=Percentage of potential
customers for EV*Percentage
Units: {%}
Percentage of non-EV drivers"= Percentage of potential
customers for EV*(total number of vehicle-(total number of
HEV +total number of BEV))/total number of vehicle
Units: Dmnl
"Percentage of potential non-EV customers"=1-Percentage of
potential customers for EV
Units: Dmnl
planned construction=Gap/Planning delay
Units: station/year
Population in Istanbul= WITH LOOKUP (Time, ([(2010,0)(2050,2.2e+007)],
(2010,1.4e+007),(2013.75,1.50526e+007),(2019.1,1.64561e+007),
125
(2025.63,1.79649e+007),(2031.62,1.90
77e+0
7),
(2039.54,2.03158e+007),(2045,2.1e+007),(2050,2.1e+007) ))
Units: person
Reduction of CO2= (Total emission level when all cars are CV
- Total emission level)/Total emission level when all cars
are CV
Units: Dmnl
"Smoothed learning ratio for R&D"=
DELAY1("EV learning ratio for R&D", "Time delay for R&D" )
Units: Dmnl
"Social exposure of non-EV drivers"=
"Effectiveness
of
word
of
mouth
of
non-EV
drivers"*"Percentage of non-EV drivers"
Units: Dmnl/year
"Total BEV sales (Total experience in Istanbul)"= INTEG (EV
sale accumulation,20)
Units: vehicle
Total emission level=
Emission level of total CVs+Emission level of total
BEVs+Emission level of total HEVs
Units: g/year
Total emission level when all cars are CV= Annual range of CV
*Emission level of one CV*(total number of
CV
+
total
number of BEV + total number of HEV)
Units: g/year
"Total HEV sales in Istanbul (total experience)"= INTEG (HEV
sale accumulation,100)
Units: vehicle
total market share=CV sales market share+BEV sales market
share+HEV sales market share
Units: Dmnl
total number of BEV= INTEG (BEV sales-BEV discard rate,20)
Units: vehicle
total number of HEV= INTEG (HEV sales-HEV discard rate,100)
Units: vehicle
total
number
of
CV=
INTEG
(CV
sales-CV
discard
rate,1.99506e+006)
Units: vehicle
total number of vehicle= INTEG (market growth,1.99506 e+006)
Units: vehicle
"Total re-purchase"=CV discard rate+BEV discard rate+HEV
discard rate
Units: vehicle/year
Total social exposure= Marketing influence on EV+Social
exposure of EV drivers+
"Social exposure of non-EV drivers"
Units: Dmnl/year
Total vehicle demand=motorization rate*Population in Istanbul
Units: vehicle
126
APPENDIX B: PARAMETER VALUES
Table B.1. Parameter values.
Parameter
Alpha for BEV driving range
Alpha for BEV purchase price
Alpha for BEV refueling time
Alpha for CV emission rate
Alpha for CV purchase price
Annual range of BEV
Annual range of CV
Annual range of HEV
BEV allocated money per car for B
BEV annual income level of household
for A
BEV discard fraction
BEV emission rate
BEV estimation time
BEV maintenance cost
BEV private consumption tax
BEV reference experience
BEV reference experience R&D
BEV weight of emission utility for A
BEV weight of emission utility for B
BEV weight of maintenance cost for A
BEV weight of maintenance cost for B
BEV weight of purchase price A
BEV weight of purchase price B
BEV weight of refueling cost for A
BEV weight of refueling cost for B
BEV weight of time utility for A
BEV weight of time utility for B
Construction delay
Current electricity unit price
CV allocated money per car for B
CV annual income level of household
for A
CV discard fraction
CV driving range
CV number of refueling point in every
station
CV private consumption tax
CV refueling time
CV weight of emission utility for A
CV weight of emission utility for B
Value of parameter
0,023
-0,015
-0,015
0
-0,029
18000
18000
18000
80000
60000
Unit of Parameter
Dmnl
Dmnl
Dmnl
Dmnl
Dmnl
km/year
km/year
km/year
Dmnl
TL
0,08
0,386
2
1
0,03
5000
5000
-0,07
-0,09
-0,085
-0,12
-0,4
-0,42
-0,17
-0,11
-0,17
-0,155
1
0,11262
80000
60000
1/year
Dmnl
Year
Dmnl
Dmnl
Vehicle
Vehicle
Dmnl
Dmnl
Dmnl
Dmnl
1/TL
1/TL
km/TL
km/TL
1/hour
1/hour
year
TL/km
TL
TL
0,08
700
8
1/year
Km
Point
0,37
1/12
-0,07
-0,09
Dmnl
Hour
Dmnl
Dmnl
127
Table B.1. Parameter values (cont).
CV weight of maintenance cost for A
CV weight of maintenance cost for B
CV weight of purchase price for A
CV weight of purchase price for B
CV weight of refueling cost for A
CV weight of refueling cost for B
CV weight of time utility for A
CV weight of time for utility B
Desired station per vehicle ratio
Effectiveness of word of mouth of EV
drivers
Effectiveness of word of mouth of nonEV drivers
Emission level of one BEV
Emission level of one CV
EV awareness loss fraction
HEV allocated money per car for B
HEV alpha
HEV annual income level of household
for A
HEV discard fraction
HEV estimation time
HEV improvement
HEV initial experience
HEV maintenance cost
HEV private consumption tax
HEV weight of emission utility for A
HEV weight of emission utility for B
HEV weight of maintenance cost for B
HEV weight of maintenance cost for B
HEV weight of refueling cost for A
HEV weight of refueling cost for B
HEV weight of purchase price for A
HEV weight of purchase price for B
HEV weight of time for A
HEV weight of time for B
marketing influence on EV
Max range
Municipality criteria
motorization rate
Number of refueling stations of CV
percentage of group A
percentage of group B
Percentage of potential customers CV
Planning delay
Reference CV nbr per refue. p.point
-0,085
-0,12
-0,4
-0,42
-0,17
-0,11
-0,17
-0,155
0,05
0,25
Dmnl
Dmnl
1/TL
1/TL
km/TL
km/TL
1/hour
1/hour
Station/vehicle
Dmnl/year
0,15
Dmnl/year
72,56
188
0,01
80000
-0.015
60000
g/(km*vehicle)
g/(km*vehicle)
Year
Dmnl
Dmnl
Dmnl
0,08
2
0,25
50000
0,5
0,37
-0,085
-0,09
-0,09
-0,12
-0,17
-0,11
-0,4
-0,42
-0,17
-0,155
0,01
300
1000
0,145
1000
0,88
0,12
1
2
240
1/year
Year
Dmnl
Vehicle
Dmnl
Dmnl
Dmnl
Dmnl
Dmnl
Dmnl
km/TL
km/TL
1/TL
1/TL
1/hour
1/hour
Dmnl/year
Km
Station/year
Vehicle/person
Station
Dmnl
Dmnl
Dmnl
year
Vehicle/(station*point)
128
Table B.1. Parameter values (cont).
Reference vehicle number
per station
time delay
Time delay for R&D
20
Vehicle/station
1
2
Year
Year
129
APPENDIX C: SENSITIVITY RESULTS
Table C.1. Final values of key variables in Sensitivity Experiments.
Base run value
(at 2042)
Sensitivity Results
(2042)
(min, max)
Sensitivity parameter: Motorization rate (vehicle/person)
Sensitivity range (minimum value, base run value, maximum value) : (0.13,
0.145,0.17)
BEV fleet market share (%)
19.76
(19.67, 19.77)
HEV fleet market share (%)
20.77
(20.74, 20.9)
CV fleet market share (%)
59.47
(59.6, 59.4)
Sensitivity parameter: Effectiveness of WoM of EV drivers (dmnl/year)
Sensitivity range (minimum value, base run value, maximum value): (0.2,0.25,0.3)
19.76
(19.07, 20.24)
20.77
(20.11, 21.42)
59.47
(60.82, 58.34)
Sensitivity parameter: Effectiveness of WoM of non- EV drivers (dmnl/year)
Sensitivity range (minimum value, base run value, maximum value): (0.12,0.15,0.18)
(18.04, 21.03)
19.76
20.77
(18.97, 22.32)
59.47
(63, 56.65)
Sensitivity parameter: Maximum range (km)
Sensitivity range (minimum value, base run value, maximum value): (240,300,360)
19.76
(20.43, 18.96)
(20.48, 21.11)
20.77
59.47
(59.09, 59.93)
Sensitivity parameter: Estimation time of vehicles (year)
Sensitivity range (minimum value, base run value, maximum value): (1.6,2,2.4)
19.76
(19.77, 19.61)
(20.78, 20.83)
20.77
59.47
(59.45, 59.56)
Sensitivity parameter: Annual range of vehicles (km/year)
Sensitivity range (minimum value, base run value, maximum value):
(14400,18000,21600)
CO2 reduction (%)
17.3
(17.3, 17.3)
Sensitivity parameters: Weight of emission utility for A (dmnl)& Weight of emission
utility for B (dmnl)
(-0.084,-0.07,-0.056) for A market segment
(-0.108,-0.09,-0.07) for B market segment
(19.6, 19.8)
19.76
20.77
(20.8,20.8)
(59.6,59.4)
59.47
130
Table C.1. Final values of key variables in Sensitivity Experiments (cont).
Sensitivity parameters: Weight of time utility for A & Weight of time utility for B
(-0.20,-0.17,-0.14)for A market segment
(-0.185,-0.155,-0.125)for B market segment
(19.04,20.35)
19.76
20.77
(20.52, 21.08)
59.47
(60.44,58.57)
Sensitivity parameters: Weight of purchase price utility for A & Weight of purchase
price utility for B (1/TL)
(-0.5,-0.4-0.3) for A market segment
(-0.51,-0.42-0.34) for B market segment
(19.6,19.77)
19.76
20.77
(20.39, 21.21)
(60.01,59.02)
59.47
Sensitivity parameters: Weight of maintenance cost utility for A & Weight of
maintenance cost utility for B (dmnl)
(-0.1,-0.085-0.07)for A market segment
(-0.144,-0.12-0.096)for B market segment
(19.55, 19.82)
19.76
20.77
(20.78, 20.82)
(59.67, 59.36)
59.47
Sensitivity parameters: Weight of refueling cost utility for A & Weight of refueling cost
utility for B (km/TL)
(-0.20,-0.17,-0.136) for A market segment
(-0.132, -0.11,-0.088) for B market segment
(19.22, 20.11)
19.76
20.77
(20.8,20.81)
(59.98, 59.08)
59.47
Sensitivity parameters: Time delay for R&D (year)
Sensitivity range (minimum value, base run value, maximum value): (1.6, 2,2.4)
(19.68, 19.69)
19.76
20.77
(20.8,20.8)
59.47
(59.5, 59.51)
Sensitivity parameters: Annual income level of household & Allocated money per car
(48000, 60000,72000) for A market segment
(64000,80000,96000) for B market segment
19.76
(19.68,19.69)
(20.77, 20.83)
20.77
59.47
(59.55,59.48)
131
Table C.1. Final values of key variables in Sensitivity Experiments (cont).
Sensitivity parameters: Reference vehicle number (vehicle/station)
Sensitivity range (minimum value, base run value, maximum value): (16, 20,24)
19.76
(19.27,19.96)
20.77
(20.96,20.7)
59.47
(59.77,59.34)
Sensitivity parameters: Marketing influence (dmnl)
Sensitivity range (minimum value, base run value, maximum value): (0.008,
0.01,0.012)
(18.76,20.42)
19.76
20.77
(19.8, 21.6)
(61.44, -57.98)
59.47
Sensitivity parameters: Discard fraction (1/year)
Sensitivity range (minimum value, base run value, maximum value): (0.064, 0.08,0.96)
(18.31, 31.56)
19.76
20.77
(19.44,31.42)
(62.25,37.02)
59.47
Sensitivity parameters: Municipality criteria
Sensitivity range (minimum value, base run value, maximum value): (800, 1000,1200)
19.76
(19.55, 19.77)
20.77
(20.87,20.76)
59.47
(59.58, 59.47)
Sensitivity parameters: Planning delay (year)
Sensitivity range (minimum value, base run value, maximum value): (1.6, 2,2.4)
19.76
(19.76,19.61)
(20.78,20.83)
20.77
59.47
(59.46,59.56)
Sensitivity parameters: Construction delay
Sensitivity range (minimum value, base run value, maximum value): (1, 1,1.2)
19.76
(19.7,19.56)
(20.8,20.84)
20.77
59.47
(59.5,59.6)
.
132
REFERENCES
1.
Gupta, H., Fundamentals of Internal Combustion Engines, Prentice Hall of India,
New Delhi, 2006.
2.
Lindly, J. K. and T. Haskew, “Impact of Electric Vehicles on Electric Power
Generation and Global Environmental Change,” Advances in Environmental
Research, Vol. 6, No. 3, pp. 291–302, 2002.
3.
EPA,
“Overview
of
Greenhouse
Gases”,
2011,
http://www.epa.gov/climatechange/ghgemissions/gases.html, accessed in May 2013.
4.
Fuglestvedt, J., T. Berntsen, G. Myhre, K. Rypdal, and R. B. Skeie, “Climate
Forcing From the Transport Sectors,” Proceedings of the National Academy of
Sciences of the United States of America, Vol. 105, No. 2, pp. 454–458, 2008.
5.
Greene, D. L., J. L. Hopson, and J. Li, “Have We Run Out of Oil Yet? Oil Peaking
Analysis From an Optimist‟s Perspective,” Energy Policy, Vol. 34, No. 5, pp. 515–
531, 2006.
6.
Aftabuzzaman, M. and E. Mazloumi, “Achieving Sustainable Urban Transport
Mobility in Post Peak Oil Era,” Transport Policy, Vol. 18, No. 5, pp. 695–702,
2011.
7.
Aldrich, B., ABC’s of Afv's: A Guide to Alternative Fuel Vehicles, 2nd Edition,
California Energy Commision, California, p. 1, 1995.
8.
Eaves, S. and J. Eaves, “A Cost Comparison of Fuel-Cell and Battery Electric
Vehicles,” Journal of Power Sources, Vol. 130, No. 1–2, pp. 208–212, 2004.
9.
Hackbarth, A. and R. Madlener, “Consumer Preferences for Alternative Fuel
Vehicles: A Discrete Choice Analysis,” Transportation Research Part D: Transport
and Environment, Vol. 25, pp. 5–17, 2013.
10.
Van Rijnsoever, F. J., P. Hagen, and M. Willems, “Preferences for Alternative Fuel
Vehicles by Dutch Local Governments,” Transportation Research Part D:
Transport and Environment, Vol. 20, No. June 2012, pp. 15–20, 2013.
11.
OECD, “Reducing Transport GHG Emissio ns- Trends and Data 2010”, 2010,
http://www.internationaltransportforum.org/Pub/pdf/10GHGTrends.pdf, accessed in
May 2013.
12.
Geyer, R., J. Koehn, and T. Olsen, Reducing Greenhouse Gas Emissions with
Hybrid-Electric Vehicles: An Environmental and Economic Analysis, Technical
Report, Donald Bren School of Environmental Science and Management, 2005.
133
13.
Scott, M., M. Kintner-Meyer, D. Elliott, and W. Warwick, Impacts Assessment of
Plug-In Hybrid Vehicles on Electric Utilities and Regional US Power Grids,
Technical Report, Pasific Northwest National Laboratory, Richland, 2007.
14.
Samaras, C. and K. Meisterling, “Life Cycle Assessment of Greenhouse Gas
Emissions From Plug- in Hybrid Vehicles: Implications for Policy,” Environmental
Science & Technology, Vol. 42, No. 9, pp. 3170–3176, 2008.
15.
Shafiei, E., H. Thorkelsson, E. I. Ásgeirsson, B. Davidsdottir, M. Raberto, and H.
Stefansson, “An Agent-Based Modeling Approach to Predict the Evolution of
Market Share of Electric Vehicles: A Case Study From Iceland,” Technological
Forecasting and Social Change, Vol. 79, No. 9, pp. 1638–1653, 2012.
16.
Sioshansi, R. and P. Denholm, “Emissions Impacts and Benefits of Plug- in Hybrid
Electric Vehicles and Vehicle-to-Grid Services,” Environmental Science &
Technology, Vol. 43, No. 4, pp. 1199–1204, 2009.
17.
Kwon, T., “Strategic Niche Management of Alternative Fuel Vehicles: A System
Dynamics Model of the Policy Effect,” Technological Forecasting and Social
Change, Vol. 79, No. 9, pp. 1672–1680, 2012.
18.
Lopes, J. A. P., F. J. Soares, P. M. R. Almeida, P. C. Baptista, C. M. Silva, T. L.
Farias, and A. G. Architecture, “Quantification of Technical Impacts and
Environmental Benefits of Electric Vehicles Integration on Electricity Grids,”
Advanced Electromechanical Motion Systems & Electric Drives Joint Symposium,
pp. 1–3, 2009.
19.
Wansart, J. and E. Schnieder, “Modeling Market Development of Electric
Vehicles,” 4th Annual IEEE Systems Conference, pp. 371–376, 2010.
20.
Von Albrichsfeld, C. and J. Karner, “Brake System for Hybrid and Electric
Vehicles,” in Electric and Hybrid-Electric Braking Systems and NVH
Considerations, R. K. Jurgen, Pennsylvania, pp. 11–17, 2011.
21.
Manzie, C., H. Watson, and S. Halgamuge, “Fuel Economy Improvements for
Urban Driving: Hybrid vs. Intelligent Vehicles,” Transportation Research Part C:
Emerging Technologies, Vol. 15, No. 1, pp. 1–16, 2007.
22.
Zhang, L., T. Brown, and G. S. Samuelsen, “Fuel Reduction and Electricity
Consumption Impact of Different Charging Scenarios for Plug- in Hybrid Electric
Vehicles,” Journal of Power Sources, Vol. 196, No. 15, pp. 6559–6566, 2011.
23.
Andersson, S.-L., A. K. Elofsson, M. D. Galus, L. Göransson, S. Karlsson, F.
Johnsson, and G. Andersson, “Plug- in Hybrid Electric Vehicles as Regulating
Power Providers: Case Studies of Sweden and Germany,” Energy Policy, Vol. 38,
No. 6, pp. 2751–2762, 2010.
24.
Christidis, P., H. Hernandez, and A. Georgakaki, Hybrids for Road Transport,
Technical Report, Institute for Prospective Technological Studies, Spain, 2005.
134
25.
Høyer, K. G., “The History of Alternative Fuels in Transportation: The Case of
Electric and Hybrid Cars,” Utilities Policy, Vol. 16, No. 2, pp. 63–71, 2008.
26.
Delucchi, M. and T. Lipman, “An Analysis of the Retail and Lifecycle Cost of
Battery-Powered Electric Vehicles,” Transportation Research Part D, Vol. 6, pp.
371–404, 2001.
27.
Egbue, O. and S. Long, “Barriers to Widespread Adoption of Electric Vehicles: An
Analysis of Consumer Attitudes and Perceptions,” Energy Policy, Vol. 48, pp. 717–
729, 2012.
28.
Wiedmann, K.-P., N. Hennigs, L. Pankalla, M. Kassubek, and B. Seegebarth,
“Adoption Barriers and Resistance to Sustainable Solutions in the Automotive
Sector,” Journal of Business Research, Vol. 64, No. 11, pp. 1201–1206, 2011.
29.
TUBITAK, “TUBITAK‟ta Elektrikli Araç Toplantısı Düzenlendi”, 2012,
http://www.tubitak.gov.tr/tr/haber/tubitakta-elektrikli-araclar-toplantisi-duzenlendi,
accessed in December 2012.
30.
Chan, C. C., “The State of the Art of Electric, Hybrid, and Fuel Cell Vehicles,”
Proceedings of the IEEE, Vol. 95, No. 4, pp. 704–718, 2007.
31.
Dijk, M., R. J. Orsato, and R. Kemp, “The Emergence of an Electric Mobility
Trajectory,” Energy Policy, Vol. 52, pp. 135–145, 2013.
32.
Ehsani, M., Y. Gao, S. E. Gay, and A. Emadi, Modern Electric, Hybrid Electric, and
Fuel Cell Vehicles Fundamentals, Theory, and Design, CRC Press, Florida, pp. 99–
114, 2005.
33.
Husain, I., Electric and Hybrid Vehicles Design Fundamentals, CRC Press, Florida,
pp. 1–3, 2003.
34.
Emadi, A., S. Member, Y. J. Lee, S. Member, and K. Rajashekara, “Power
Electronics and Motor Drives in Electric , Hybrid Electric , and Plug-In Hybrid
Electric Vehicles,” Industrial Electronics, Vol. 55, No. 6, pp. 2237–2245, 2008.
35.
German, J. M., Hybrid-Powered Vehicles, 2nd Edition, SAE International,
Pennsylvania, pp. 9–14, 2011.
36.
Markel, T. and A. Simpson, “Cost-Benefit Analysis of Plug-In Hybrid Electric
Vehicle Technology,” 22nd International Electric Vehicle Symposium, Vol. 1, 2006
37.
Gonder, J. and A. Simpson, “Measuring and Reporting Fuel Economy of Plug-In
Hybrid Electric Vehicles,” 22nd International Battery, Hybrid and Fuel Cell Electric
Vehicle Symposium and Exhibition, 2006.
38.
Lemoine, D. M., D. M. Kammen, and a E. Farrell, “An Innovation and Policy
Agenda for Commercially Competitive Plug- in Hybrid Electric Vehicles,”
Environmental Research Letters, Vol. 3, No. 1, p. 1-10, 2008.
135
39.
Shepherd, S., P. Bonsall, and G. Harrison, “Factors Affecting Future Demand for
Electric Vehicles: A Model Based Study,” Transport Policy, Vol. 20, pp. 62–74,
2012.
40.
Rogers, E. M., Diffusion of Innovations, 5th edition, Free Press, New York, p. 35,
2003.
41.
Sullivan, J., I. Salmeen, and C. Simon, PHEV Marketplace Penetration: An Agent
Based Simulation, Technical Report, University of Michigan, Transportation
Research Institute, 2009.
42.
Tran, M., “Technology-Behavioural Modelling of Energy Innovation Diffusion in
the UK,” Applied Energy, Vol. 95, pp. 1–11, 2012.
43.
Struben, J. and J. D. Sterman, “Transition Challenges for Alternative Fuel Vehicle
and Transportation Systems,” Environment and Planning B: Planning and Design,
Vol. 35, No. 6, pp. 1070–1097, 2008.
44.
Hoogma, R., R. Kemp, J. Schot, and B. Truffer, Experimenting for Sustainable
Transport the Approach of Strategic Niche Management, Spon Press, London, 2002.
45.
Zhang, T., S. Gensler, and R. Garcia, “A Study of the Diffusion of Alternative Fuel
Vehicles: An Agent-Based Modeling Approach,” Journal of Product Innovation
Management, Vol. 28, No. 2, pp. 152–168, 2011.
46.
Eppstein, M. J., D. K. Grover, J. S. Marshall, and D. M. Rizzo, “An Agent-Based
Model to Study Market Penetration of Plug- in Hybrid Electric Vehicles,” Energy
Policy, Vol. 39, No. 6, pp. 3789–3802, 2011.
47.
Sterman, J. D., Business Dynamics: System Thinking and Modeling for a Complex
World, Irwin/McGraw-Hill, Boston, 2000.
48.
Barlas, Y., System Dynamics: Systemic Feedback Modeling for Policy Analysis,
UNESCO-EOLSS Publishers, Paris, pp. 1131–1175, 2002.
49.
Seker, O., Modeling the Dynamics of Thyroid Hormones and Related Disorders,
MSc Thesis, Boğaziçi University, Industrial Engineering Department, 2012.
50.
Turkish Statistical Institute, “İllere Göre Motorlu Kara Taşıtları Sayısı”, 2012,
http://www.tuik.gov.tr/PreHaberBultenleri.do?id=13610, accessed in May 2013.
51.
Turkish Statistical Institute, “N üfus Projeksiyonları, 2013-2075”, 2013,
http://www.tuik.gov.tr/PreHaberBultenleri.do?id=15844, accessed in June 2013
52.
McFadden, D., “Econometric Models for Probabilistic Choice Among Products,”
The Journal of the Business, Vol. 53, No. 3, pp. 13–29, 1980.
53.
TEB ARVAL and Corporate Vehicle Observatory, “Şirket Filolarındaki Araç Sayısı
1.8
Milyona
Ulaştı”,
2012,
http://www.tebarval.com.tr/tur/operasyonel-
136
kiralama/home/basin-odasi/basin-bultenleri/filo-barometre.html, accessed in July
2013.
54.
Egbue, O. and S. Long, “Barriers to widespread adoption of electric vehicles: An
analysis of consumer attitudes and perceptions,” Energy Policy, Vol. 48, pp. 717–
729, 2012.
55.
Davis, F. D., “Perceived Usefulness , Perceived Ease Of Use, and User Acceptance
of Information Technology,” MIS Quarterly, Vol. 13, No. 3, pp. 319–340, 1989.
56.
Brownstone, D., D. S. Bunch, and K. Train, “Joint Mixed Logit Models of Stated
and Revealed Preferences for Alternative Fuel Vehicles,” Transportation Research
Part B, Vol. 34, pp. 315–338, 2000.
57.
Yücel, G., “Facilitating Diffusion of Innovations Through Niche Management
Strategies,” 31st International System Dynamics Conference, Massachusetts, p. 116,
2013.
58.
Granovskii, M., I. Dincer, and M. A. Rosen, “Economic and Environmental
Comparison of Conventional, Hybrid, Electric and Hydrogen Fuel Cell Vehicles,”
Journal of Power Sources, Vol. 159, No. 2, pp. 1186–1193, 2006.
59.
Barut, A., “Eşarj-Şarj İstasyonları”, 2013, http://esarj.com/elektrikli-arac-sarjistasyonlari/, accessed in January 2013.
60.
Gelir Dairesi Başkanlığı, “Özel Tüketim Vergisi Tutarları ve Oranları”, 2012,
http://www.gib.gov.tr/fileadmin/mevzuatek/otv_oranlari_tum/22_09_2012_II_sayili
_liste.htm, accessed in April 2013.
61.
Tran, M., D. Banister, J. D. K. Bishop, and M. D. McCulloch, “Simulating Early
Adoption of Alternative Fuel Vehicles for Sustainability,” Technological
Forecasting and Social Change, Vol. 80, No. 5, pp. 865-875, 2013.
62.
Lemoine, D. M. and D. M. Kammen, “Addendum to „An Innovation and Policy
Agenda for Commercially Competitive Plug-in Hybrid Electric Vehicles‟,”
Environmental Research Letters, Vol. 4, No. 3, pp. 1-3, 2009.
63.
Helms, H., M. Pehnt, U. Lambrecht, and A. Liebich, “Electric Vehicle and Plug- in
Hybrid Energy Efficiency and Life Cycle Emissions,” 18th International
Symposium Transport and Air Pollution Session 3: Electro and Hybrid Vehicles, pp.
113–124, 2010.
64.
United States Enviormental Protection Agency, “Greenhouse Gas Emissions From a
Typical
Passenger
Vehicle”,
2011,
http://www.epa.gov/otaq/climate/documents/420f11041.pdf, accessed in June 2013.
65.
TEİAŞ, “Türkiye Elektrik Enerjisi Üretim Planlama Çalışması”,
http://www.teias.gov.tr/ElektrikUretimPlani.aspx, accessed in January 2013.
2004,
137
66.
Environmental Protection Agency, “Average Annual Emissions and Fuel
Consumption for Gasoline-Fueled Passenger Cars and Light Trucks”, 2008,
http://www.epa.gov/otaq/climate/documents/420f08024.pdf, accessed in May 2013.
67.
Barlas, Y., “Formal Aspects of Model Validity and Validation in System
Dynamics,” System Dynamics Review, Vol. 12, No. 3, pp. 183–210, 1996.
68.
World Energy Council Turkish National Committe, Enerji Raporu, Poyraz Ofset,
Ankara, 2011.

ANALYZING ELECTRIC VEHICLE DIFFUSION SCENARIOS FOR

Transkript

Benzer belgeler

Concept Note Call Acronym: H2020-GV-5-2014 (Single

VEHICLE FOR SALE VOLVO S 80 2.5 T

EVS Projects in Switzerland

important notice on the vehicle inspection of vehicles with

Yuvarlanma Direnci

Connected Vehicle

Academic @ Paper Possibility of Turkey to Transit Electric Vehicle

EVS Opportunity in Youmore www.youmore.org

Learn 500 useful words and expressions for travelling